RouterOS v5.9 released

honzam · Tue Nov 29, 2011 6:37 pm

What's new in 5.9 (2011-Nov-29 14:32):

*) ssh - fix memory leak when client uses public key authentication;
*) ppp - added support for new RADIUS attribute MT-Delegated-IPv6-Pool (#22);
*) ntp client - faster initial synchronization;
*) ppp - added support for dhcpv6 pd;
*) wireless - nv2 improvements for 11n cards;
*) hotspot - fixed login page to better handle big load;
*) wireless - change default rate-selection to advanced;
*) snmp - fixed simple queue table;
*) webfig - fixed problem were users without sensitive permission could download
sensitive files (like backups);
*) webfig - fixed problem were table filters did not work always as expected;
*) metarouter - fixed problem where local routeros instances did not boot;
*) dhcpv6 - client and server moved to respective /ipv6 dhcp- entry;
*) dhcpv6 server - changed how bindings are defined, users should add
missing static binding information after upgrade;
*) sms - send sms now uses channel from config if it's not specified in the command;

honzam · Tue Nov 29, 2011 7:19 pm

For rb4xx new BIOS 2.38. What changes?

omega-00 · Tue Nov 29, 2011 10:11 pm

Anyone got a dual-stacked IPv4/IPv6 PPPoE connection up and running yet?

I'm playing around with mine now, so if anyone has a working configuration please post a copy (minus auth details obviously ;-D).

Insider · Wed Nov 30, 2011 12:40 am

No more wireless changes yet?? Looking forward xmas eve present, oink oink

Have to say Mikrotik is moving forward much better then we have had seen in last year backview.. Thumbs up.. This way we gonna announce IPV6 implementation soon to our custs..

Djelibeybi · Wed Nov 30, 2011 12:47 am

Anyone got a dual-stacked IPv4/IPv6 PPPoE connection up and running yet?

Nope.

I've added the dhcp-client entry, but it's just stuck on "Searching...". I suspect a firewall issue somewhere.

oeyre · Wed Nov 30, 2011 12:51 am

Any progress with #2011062966000309 or #2011111766000189 ?

omega-00 · Wed Nov 30, 2011 2:48 am

Nope. I've added the dhcp-client entry, but it's just stuck on "Searching...". I suspect a firewall issue somewhere.

I get exactly the same, just sits on "Searching..."

Djelibeybi · Wed Nov 30, 2011 2:55 am

I get exactly the same, just sits on "Searching..."

If I open IPv6 outbound udp/567 and inbound udp/546 I can see packets leaving the RouterOS, but nothing coming back. I know that DHCPv6-PD works, as I've tested it with a generic Linux-based router and a Fritz!Box device. I'm not sure why RouterOS is not getting the returned packets.

jcem · Wed Nov 30, 2011 3:00 am

When RB reaches high load (100%) for a a while it reboots. I had to downgrade to 5.8 again.

I even had to disable some of the wireless cards to be able to upload the 5.8 version so it wouldn´t reboot while
uploading.

I could not rebuild the Userman database without the RB rebooted itself

SUPOUT sent to MT

WirelessRudy · Wed Nov 30, 2011 3:26 am

What's new in 5.9 (2011-Nov-29 14:32):

*) ssh - fix mempry leak when client uses public key authentication;
*) ppp - added support for new RADIUS attribute MT-Delegated-IPv6-Pool (#22);
*) ntp client - faster initial synchronization;
*) ppp - added support for dhcpv6 pd;
*) wireless - nv2 improvements for 11n cards;
*) hotspot - fixed login page to better handle big load;
*) wireless - change default rate-selection to advanced;
*) snmp - fix simple queue table;
*) webfig - fixed problem were users wihtout sensitive permission could download
senstive files (like backups);
*) webfig - fixed problem were table filters did not work allways as expected;
*) metarouter - fixed problem where local routeros instances did not boot;
*) dhcpv6 - client and server moved to respective /ipv6 dhcp- entry;
*) dhcpv6 server - changed how bindings are defined, users should add
missing static binding information after upgrade;
*) sms - send sms now uses channel from config if it's not specified in the command;

who is "honzam"? Is this the official MT release of v5.9? How official is this release?

WirelessRudy · Wed Nov 30, 2011 3:28 am

No more wireless changes yet?? Looking forward xmas eve present, oink oink

I see two wireless changes/improvement in the changelog. What is it you are waiting for?

Wed Nov 30, 2011 7:55 am

do you need more official than the official webpage?
http://www.mikrotik.com/download.html

ilius168 · Wed Nov 30, 2011 8:30 am

- Locked up RB 750GL - unable to remote to the router.

- Lights up blinking on all ethernet ports.

- unable to reset

Ready to dismantle the Router and to be sent for RMA

Wed Nov 30, 2011 8:36 am

- Locked up RB 750GL - unable to remote to the router.

- Lights up blinking on all ethernet ports.

- unable to reset

Ready to dismantle the Router and to be sent for RMA

this is definitely not related to v5.9 or any other RouterOS version. you got yourself a hardware issue, contact support

Wed Nov 30, 2011 8:37 am

can you all please test the "Latvia" download mirror? What speeds to you get in your country (name it). Is it higher than ususal?

alexspils · Wed Nov 30, 2011 8:51 am

Latvia.100 mbps

omega-00 · Wed Nov 30, 2011 9:07 am

Speed is same if not slower than the regular latvia mirror (most mirrors except the Australian ones are slow for us due to RTT affect on TCP).

mikrotik-cloud.PNG

Wed Nov 30, 2011 9:09 am

Speed is same if not slower than the regular latvia mirror (most mirrors except the Australian ones are slow for us due to RTT affect on TCP).

mikrotik-cloud.PNG

you are the first in your country to download. the second person should get it quickly.

honzam · Wed Nov 30, 2011 9:10 am

who is "honzam"? Is this the official MT release of v5.9? How official is this release?

No, I'm a member of the forum. You are not enough official download page?

Wed Nov 30, 2011 9:42 am

Anyone got a dual-stacked IPv4/IPv6 PPPoE connection up and running yet?
Nope. I've added the dhcp-client entry, but it's just stuck on "Searching...". I suspect a firewall issue somewhere.

Here is the working configuration example:
http://wiki.mikrotik.com/wiki/Manual:IPv6_PD_over_PPP

If you have any problems send a supout to support.

Wed Nov 30, 2011 9:50 am

I am not having any luck with DHCPv6-PD client via PPP and my setup matches the wiki article.

karina · Wed Nov 30, 2011 9:52 am

download to Spain from Latvia 1.2Mbps. Much faster than normal.

honzam · Wed Nov 30, 2011 9:56 am

download to Czech Republic - 1,1MB realy FAST

oeyre · Wed Nov 30, 2011 10:02 am

who is "honzam"? Is this the official MT release of v5.9? How official is this release?
No, I'm a member of the forum. You are not enough official download page?

I believe the concern was that regular users were announcing software releases instead of MikroTik staff. Additionally it was being done before the usual email from MikroTik was sent to notify everyone about a software release.

Wed Nov 30, 2011 10:03 am

this is a user forum. users post everything and are the main driving force here.

janisk · Wed Nov 30, 2011 10:05 am

I am not having any luck with DHCPv6-PD client via PPP and my setup matches the wiki article.

what happens?

Also, check on dhcp-pd server side (if it is ROS) that you have dynamic /ipv6 dhcp-server added to the interface. And if there is problem with that, usually ppp session is terminated. So, in ppp profile you have to set ipv6 prefix that has to be used. After that, dhcp-pd client should get pool instantly. Note - client will add pool, not set up address on some interface.

janisk · Wed Nov 30, 2011 10:07 am

I believe the concern was that regular users were announcing software releases instead of MikroTik staff. Additionally it was being done before the usual email from MikroTik was sent to notify everyone about a software release.

it is allways possible to go to download page and check, what version is available from there. And if check history of releases - there are a lot of release topics made by users if we where slow to create one.

Djelibeybi · Wed Nov 30, 2011 10:20 am

Here is the working configuration example:
http://wiki.mikrotik.com/wiki/Manual:IPv6_PD_over_PPP

This doesn't work -- my configuration is identical and my RB client never received a prefix nor created the IPv6 pool. This works with other routers, e.g Linux running wide-dhcpv6-client on the same PPPoE connection. The other side is my ISP and not a RouterBoard. I will email support now.

WirelessRudy · Wed Nov 30, 2011 11:01 am

do you need more official than the official webpage?
http://www.mikrotik.com/download.html

Last time I complained about the fact I got my alert e-mail only 5 days after the release and now new upgrades are not even published on this forum any more? The download page of the website is not the usual page I start my day with. And by seeing a unknown name publishing the release in the forum made me a bit weary about the official status....
I'm not the type of guy that opens the e-mail with a Pamela Anderson picture promising you a good experiance after which you'll find your hd erased or worse.
So if a release is published by someone I have never heard of before I am a bit confused yes....
But I'll guess in the future I have to make the download page the default opening page in my browser.. how exited the world can be..

mwheeler · Wed Nov 30, 2011 11:27 am

I'm having some issues with IPv6 over PPP. I put require IPv6 in the profile, PPP connects however DHCP client never joins. This is on Internode http://ipv6.internode.on.net/ . I'm yet to test with a non mtik router, because, well, that takes effort. Any sort of IPv6 DHCP logging I can enable to see what's going on?

Good work on starting to support it though

n21roadie · Wed Nov 30, 2011 11:30 am

*) wireless - nv2 improvements for 11n cards;

This probally explains why there was no throughput improvement on a PTP link using XR5 cards on a 411ah boards, has anybody noted improvements with 11n cards?

brainy · Wed Nov 30, 2011 11:43 am

I have a PTMP link with OmniTIK and SXT over a distance of 2.5km. The signal is 5dB better now (with 5.8 it was -75 now it is -70) which is nice but the throughput is still the same as before (i am using "long" as guard-interval as this was more stable/better in my last tests).

So whats the advantage?

Regards,
Joerg

Djelibeybi · Wed Nov 30, 2011 12:27 pm

I'm having some issues with IPv6 over PPP. I put require IPv6 in the profile, PPP connects however DHCP client never joins. This is on Internode http://ipv6.internode.on.net/ .

Yeah, I'm with Internode as well, same problem. I've already emailed support and will enable DHCP debug logging and send that to them. I'm waiting for them to email me the steps on how to do this.

wispwest · Wed Nov 30, 2011 12:31 pm

why is it my Nstreme link performances get worse and worse as new firmwares come out. Back to 4.17 for me and everything is top notch. 5.9 my Nstreme PtP links take a crap in data rates and performance. NV2 works good yes, but I NEED that lower latency that only Nstreme can provide.

omega-00 · Wed Nov 30, 2011 12:51 pm

I'm having some issues with IPv6 over PPP. I put require IPv6 in the profile, PPP connects however DHCP client never joins. This is on Internode http://ipv6.internode.on.net/ .
Yeah, I'm with Internode as well, same problem. I've already emailed support and will enable DHCP debug logging and send that to them. I'm waiting for them to email me the steps on how to do this.

That makes 3 of us. Please let us know how you go.

janisk · Wed Nov 30, 2011 1:03 pm

about IPv6 dhcp-pd

check what pool you have set up and what prefix you are giving out

here is how it works for me:
on server side set up pool and add that pool to the profile:

/ipv6 pool add name=poo prefix=2001:db8:beef::/48 prefix-length=62
/ppp profile set default-encryption dhcpv6-pd-pool=ipv6pool

on client:

/interface pppoe-client add allow=pap,chap,mschap1,mschap2 disabled=no interface=ether2-Out max-mru=1480 \ 
    max-mtu=1480 mrru=disabled name=pppoe-out1 password=uuu profile=default-encryption service-name="" use-peer-dns=no user=uuu

/ipv6 dhcp-client add disabled=no interface=pppoe-out1 pool-name=ipv6pookl pool-prefix-length=64

now when everything is working dhcp-client status is bound
and pool has an address

[admin@RB1000] /ipv6> pool print detail 
Flags: D - dynamic 
 0 D id=7 name="ipv6pookl" prefix=2001:db8:bad:1::/64 prefix-length=72 expire-time=2d23h58m51s

mramos · Wed Nov 30, 2011 1:08 pm

can you all please test the "Latvia" download mirror? What speeds to you get in your country (name it). Is it higher than ususal?

Brazil. But I have limits on my nlos 900MHz link (2.5mbps)

Have this version any difference to pre-released 5.9.npk?

Regards;

omega-00 · Wed Nov 30, 2011 1:11 pm

Hi Janis,

As I don't have access to the remote end I can only say that I am setting the DHCPv6 client up on the PPPoE client interface (as listed in the guide) and while I can see the request packet go out (as far as the router is concerned) I don't see anything back.

Currently my ADSL modem is bridged to allow the MikroTik to perform all auth so I'm going to try playing around with that and see if I can determine any issues.
As far as I'm aware though, it is passing all IPv4 and IPv6 traffic with no problems as I can see other discovered neighbours appear on the PPPoE interface and I received the ::/0 route from my upstream provider.

Wed Nov 30, 2011 1:22 pm

Brazil. But I have limits on my nlos 900MHz link (2.5mbps)

is this better than you usually had?

Have this version any difference to pre-released 5.9.npk?

depends on which one, we had ~15 pre-release builds

mramos · Wed Nov 30, 2011 1:39 pm

is this better than you usually had?

Yes. There were times where I had ~300Kbps from Latvia.

depends on which one, we had ~15 pre-release builds

Hmm. I will upgrade to the released one anyway. After the weekend!

Regards;

mwheeler · Wed Nov 30, 2011 1:42 pm

about IPv6 dhcp-pd...

I don't have access to the server. DHCPv6 client just sits in searching mode.

/interface pppoe-client
add ac-name="" add-default-route=yes allow=pap,chap,mschap1,mschap2 dial-on-demand=no disabled=no \
    interface=ether2 max-mru=1480 max-mtu=1480 mrru=disabled name=Internode password=MYPASSWORDHERE profile=\
    default service-name="" use-peer-dns=no user=mwheeler@internode.on.net

/ppp profile
set default change-tcp-mss=yes name=default only-one=default use-compression=default use-encryption=\
    default use-ipv6=required use-mpls=default use-vj-compression=default

/ipv6 dhcp-client
add disabled=no interface=Internode pool-name=NODE-PD pool-prefix-length=64

Ripper · Wed Nov 30, 2011 2:30 pm

After firmware update 750g is down ... i haven't that issue with other updates.

ffernandes · Wed Nov 30, 2011 2:32 pm

d/ling fine over here

latvia mirror

NetworkPro · Wed Nov 30, 2011 2:34 pm

can you all please test the "Latvia" download mirror? What speeds to you get in your country (name it). Is it higher than ususal?

Brazil. But I have limits on my nlos 900MHz link (2.5mbps)

Have this version any difference to pre-released 5.9.npk?

Regards;

http://download2.mikrotik.com/mikrotik-5.9.iso (cloudfront through local_ISP<->b-ix<->decix<->amazon) 3.2Mbytes/s and the download was done. If it was a longer download, TCP/IP could have negotiated even more.

http://download.mikrotik.com/mikrotik-5.9.iso (US softlayer through local_ISP<->cogentco<->networklayer) 900KBytes/s

NetworkPro · Wed Nov 30, 2011 2:41 pm

I downloaded it again and it went up to 1.1MBytes but avg was around 900kbytes. Should we complain to cogentco for speed outside Europe?

Wed Nov 30, 2011 2:43 pm

I downloaded it again and it went up to 1.1MBytes but avg was around 900kbytes. Should we complain to cogentco for speed outside Europe?

sorry I misunderstood your comparison. of course download.mikrotik.com will be slower, it's in USA. you need to test download2.mikrotik.com

we plan to get rid of download.mikrotik.com if the download2.mikrotik.com is good for everybody

coffeecoco · Wed Nov 30, 2011 2:53 pm

I keep refreshing this page, just with a little hope that ppp ipv6 is functioning, but i think its def broke ? anyone else?
Im with Internode, like the other guys here.

Native 'dual-stack' IPv4 and IPv6 enabled :
When you login, you will be assigned, via a dual-stack IPv4/IPv6 PPP session:
- Your existing IPv4 address (if static) and any existing framed route(s)
- A dynamic /64 IPv6 prefix for your PPP session
- A static /56 IPv6 prefix for your LAN (if you are using a router with Prefix Delegation)

otgooneo · Wed Nov 30, 2011 2:58 pm

About download speed same as before from Mongolia. No development for RB1200 IPSEC (#2011070666000118)?

Wed Nov 30, 2011 2:59 pm

you yourself have not responded to the last question in 2011070666000118, let us know the latest information you have after following suggestions by support

Jeremy23 · Wed Nov 30, 2011 3:00 pm

That makes 3 of us. Please let us know how you go.

Make that four Internode users trying to get DHCPv6-PD going

What I notice is that the router does not get a SLAAC address on the PPPoE interface. This should work whether I use DHCPv6 or not. This means that the SLAAC-based dynamic /64 that Internode provide in addition to the DHCPv6-PD-based /56 (see Internode website for details) also does not work.

I have the address fe80::7/64 on my PPPoE interface. It is marked as dynamic, but I do not know how it decided on using that address.

For reference, my OpenWrt box gets a /64 SLAAC address over PPPoE, gets a dynamic /64 global address via SLAAC, and the static /56 DHCPv6-PD address from wide-dhcpv6-client works perfectly.

Perhaps one of us needs to provide a layer-2 VPN to a MikroTik developer to allow them to perform tests against an Internode connection.

coffeecoco · Wed Nov 30, 2011 3:17 pm

Your on to something buddy!

Wed Nov 30, 2011 3:27 pm

pppoe client does not add IPv6 address to the interface. This feature is not yet added.

DHCPv6-PD should work anyway. It would be useful to know what device is used as server on ISP side and get some logs on why server is not responding at all.

coffeecoco · Wed Nov 30, 2011 3:33 pm

Not what you asked for but maybe help?

Wed Nov 30, 2011 3:35 pm

inncom, why not just copy logs in plaintext form from the terminal? also, MRZ implied that we need logs from the ISP side too ...

coffeecoco · Wed Nov 30, 2011 3:37 pm

inncom, why not just copy logs in plaintext form from the terminal? also, MRZ implied that we need logs from the ISP side too ...

I know... so you do want to see the plain text log, will it help ?

coffeecoco · Wed Nov 30, 2011 3:39 pm

Here anyway for what its worth

23:37:59 pppoe,ppp,debug,packet     <name dano1980@ipv6.internode.on.net> 
23:38:02 pppoe,ppp,debug,packet  Internode DYN: sent CHAP Response id=0x45 
23:38:02 pppoe,ppp,debug,packet     <response len=10> 
23:38:02 pppoe,ppp,debug,packet     <name dano1980@ipv6.internode.on.net> 
23:38:05 system,info,account user admin logged in from 192.168.10.39 via telnet 
23:38:05 pppoe,ppp,debug,packet  Internode DYN: sent CHAP Response id=0x45 
23:38:05 pppoe,ppp,debug,packet     <response len=10> 
23:38:05 pppoe,ppp,debug,packet     <name dano1980@ipv6.internode.on.net> 
23:38:07 pppoe,ppp,debug,packet  Internode DYN: rcvd CHAP Success id=0x45 
23:38:07 pppoe,ppp,info Internode DYN: authenticated 
23:38:07 pppoe,ppp,debug Internode DYN: IPCP lowerup 
23:38:07 pppoe,ppp,debug,packet  Internode DYN: sent IPCP ConfReq id=0x5 
23:38:07 pppoe,ppp,debug,packet     <addr 0.0.0.0> 
23:38:07 pppoe,ppp,debug,packet     <ms-dns 0.0.0.0> 
23:38:07 pppoe,ppp,debug,packet     <ms-dns 0.0.0.0> 
23:38:07 pppoe,ppp,debug Internode DYN: IPCP open 
23:38:07 pppoe,ppp,debug Internode DYN: IPV6CP lowerup 
23:38:07 pppoe,ppp,debug,packet  Internode DYN: sent IPV6CP ConfReq id=0x3 
23:38:07 pppoe,ppp,debug,packet     <interface-identifier 0:0:0:32> 
23:38:07 pppoe,ppp,debug Internode DYN: IPV6CP open 
23:38:07 pppoe,ppp,debug Internode DYN: MPLSCP lowerup 
23:38:07 pppoe,ppp,debug,packet  Internode DYN: sent MPLSCP ConfReq id=0x3 
23:38:07 pppoe,ppp,debug Internode DYN: MPLSCP open 
23:38:07 pppoe,ppp,debug Internode DYN: BCP lowerup 
23:38:07 pppoe,ppp,debug,packet  Internode DYN: sent BCP ConfReq id=0x3 
23:38:07 pppoe,ppp,debug,packet    <mac-support 1> 
23:38:07 pppoe,ppp,debug,packet    <802-tagged-frame 1> 
23:38:07 pppoe,ppp,debug,packet    <management-inline> 
23:38:07 pppoe,ppp,debug Internode DYN: BCP open 
23:38:07 pppoe,ppp,debug Internode DYN: CCP lowerup 
23:38:07 pppoe,ppp,debug Internode DYN: CCP open 
23:38:07 pppoe,ppp,debug,packet  Internode DYN: rcvd IPCP ConfReq id=0x1 
23:38:07 pppoe,ppp,debug,packet     <addr 150.101.180.143> 
23:38:07 pppoe,ppp,debug,packet  Internode DYN: sent IPCP ConfAck id=0x1 
23:38:07 pppoe,ppp,debug,packet     <addr 150.101.180.143> 
23:38:07 pppoe,ppp,debug,packet  Internode DYN: rcvd IPCP ConfNak id=0x5 
23:38:07 pppoe,ppp,debug,packet     <addr 59.167.154.243> 
23:38:07 pppoe,ppp,debug,packet     <ms-dns 192.231.203.150> 
23:38:07 pppoe,ppp,debug,packet     <ms-dns 192.231.203.151> 
23:38:07 pppoe,ppp,debug,packet  Internode DYN: sent IPCP ConfReq id=0x6 
23:38:07 pppoe,ppp,debug,packet     <addr 59.167.154.243> 
23:38:07 pppoe,ppp,debug,packet     <ms-dns 192.231.203.150> 
23:38:07 pppoe,ppp,debug,packet     <ms-dns 192.231.203.151> 
23:38:07 pppoe,ppp,debug,packet  Internode DYN: rcvd IPV6CP ConfReq id=0x1 
23:38:07 pppoe,ppp,debug,packet     <interface-identifier 224:14ff:fe9a:bc00> 
23:38:07 pppoe,ppp,debug,packet  Internode DYN: sent IPV6CP ConfAck id=0x1 
23:38:07 pppoe,ppp,debug,packet     <interface-identifier 224:14ff:fe9a:bc00> 
23:38:07 pppoe,ppp,debug,packet  Internode DYN: rcvd IPV6CP ConfAck id=0x3 
23:38:07 pppoe,ppp,debug,packet     <interface-identifier 0:0:0:32> 
23:38:07 pppoe,ppp,debug Internode DYN: IPV6CP opened 
23:38:07 pppoe,ppp,info Internode DYN: connected 
23:38:07 pppoe,ppp,debug,packet  Internode DYN: rcvd LCP ProtRej id=0x2 
23:38:07 pppoe,ppp,debug,packet      82 81 01 03 00 06 
23:38:07 pppoe,ppp,debug,packet  Internode DYN: rcvd LCP ProtRej id=0x3 
23:38:07 pppoe,ppp,debug,packet      80 31 01 03 00 0e 03 03 01 08 03 01 09 02 
23:38:07 pppoe,ppp,debug,packet  Internode DYN: rcvd IPCP ConfAck id=0x6 
23:38:07 pppoe,ppp,debug,packet     <addr 59.167.154.243> 
23:38:07 pppoe,ppp,debug,packet     <ms-dns 192.231.203.150> 
23:38:07 pppoe,ppp,debug,packet     <ms-dns 192.231.203.151> 
23:38:07 pppoe,ppp,debug Internode DYN: IPCP opened 
23:38:07 system,info dns changed

NetworkPro · Wed Nov 30, 2011 3:47 pm

We can find out the brand from the MAC address of the AC (a sniff or a PPPoE Scan reveals it).

otgooneo · Wed Nov 30, 2011 4:01 pm

you yourself have not responded to the last question in 2011070666000118, let us know the latest information you have after following suggestions by support

Yes It is my fault. But the last confirmation was not about RB1200 IPSEC throughput and latency issue. Support question was about input failure issue. Anyway I replied to support.

WirelessRudy · Wed Nov 30, 2011 5:46 pm

this is a user forum. users post everything and are the main driving force here.

Yes, but one of you guys ALWAYS announced the new upgrade on the forum first together with the change log.
If this now is done by someone I never heard of before and not even from Latvia makes small alarm bells start to move (they don't rinkle yet). Is this a hoax or even somebody just beeing funny?
I would really appreciate it if announcements of official releases are done by the source, MT. Not by some users that happen to have found some of the information.

NetworkPro · Wed Nov 30, 2011 6:08 pm

I trust myself to announce stuff

mikko · Wed Nov 30, 2011 6:32 pm

can you all please test the "Latvia" download mirror? What speeds to you get in your country (name it). Is it higher than ususal?

Finland, about 80 Mbps. Lightning fast compared to old mirror.

sda · Wed Nov 30, 2011 6:53 pm

*) hotspot - fixed login page to better handle big load;

Can someone please elaborate a bit more on this fix?

As almost everyone has modified the original hotspot login page to his needs.

Cheers

hajde · Wed Nov 30, 2011 8:20 pm

Latvia->Bosnia

mtmx80 · Wed Nov 30, 2011 9:48 pm

Hi,

Here is result from Hungary/Budapest:
wget http://download2.mikrotik.com/all_packa ... be-5.9.zip
100%[=============================================>] 14,358,469 9.33M/s in 1.5s
2011-11-30 19:32:22 (9.33 MB/s) - `all_packages-mipsbe-5.9.zip' saved [14358469/14358469]

Here is result from USA/Los Angeles:
wget http://download2.mikrotik.com/all_packa ... be-5.9.zip
100%[=============================================>] 14,358,469 3.80M/s in 5.7s
2011-11-30 11:37:36 (2.38 MB/s) - all_packages-mipsbe-5.9.zip.1

My only concern is that this came from Amazon Cloud (Frankfurt) and not from Latvia even if I select Latvia.
I would be very happy to see similar results while your server is behind a Mikrotik Router

can you all please test the "Latvia" download mirror? What speeds to you get in your country (name it). Is it higher than ususal?

omega-00 · Wed Nov 30, 2011 9:53 pm

pppoe client does not add IPv6 address to the interface. This feature is not yet added.

DHCPv6-PD should work anyway. It would be useful to know what device is used as server on ISP side and get some logs on why server is not responding at all.

I have access to Internode Corporate support so will give them a call today and see what information I can collect.

dbmillar · Wed Nov 30, 2011 10:05 pm

Any update on #2011110266000895 ?

Anyone else having problems with BGP VPLS dropping and having to disable and enable it to bring it back online again?

blake · Wed Nov 30, 2011 10:26 pm

My only concern is that this came from Amazon Cloud (Frankfurt) and not from Latvia even if I select Latvia.

They likely just chose the Latvian download location as a way to test Amazon distribution. I assume if it works well they will remove all of the country-specific download locations and just have one URL which is distributed via Amazon.

And for the record, my download from Amazon (IPv4) was much faster than the SoftLayer US mirror (IPv6). Since most people are seeing an increase in download speed I doubt my results are protocol related. I have native IPv6 transit.

TonyJr · Wed Nov 30, 2011 11:42 pm

R52Hn - Upgrade has improved signal strength from 54% to 75-76% compared to v5.8

mxmxmxmxmx · Thu Dec 01, 2011 2:20 am

For rb4xx new BIOS 2.38. What changes?

That, when I try to do

system ro up

my RB433AH do reboot, but not upgrade

NetworkPro · Thu Dec 01, 2011 2:41 am

mxmxmxmxmx, maybe NetInstall can help ?

fewi · Thu Dec 01, 2011 2:52 am

Dunno if you're still looking for download speed info. Here's my home connection in the Northeastern US:

sh-3.2$ curl -O http://download2.mikrotik.com/all_packages-mipsbe-5.9.zip
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100 13.6M  100 13.6M    0     0  1718k      0  0:00:08  0:00:08 --:--:-- 1736k
sh-3.2$

I have a 15Mbps Internet connection so 1718KB/s downloads are about as good as it gets.

Also tried at work - first one took a bit, the second one was blazing fast:

$ curl -O http://download2.mikrotik.com/all_packages-mipsbe-5.9.zip
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100 13.6M  100 13.6M    0     0  10.5M      0  0:00:01  0:00:01 --:--:-- 10.6M

Oddly enough that one didn't benefit from the same local mirror even though I live 2 miles from work, but it's completely different ISPs so that might account for it.

Thu Dec 01, 2011 8:23 am

They likely just chose the Latvian download location as a way to test Amazon distribution. I assume if it works well they will remove all of the country-specific download locations and just have one URL which is distributed via Amazon.

very good! exactly correct

janisk · Thu Dec 01, 2011 9:11 am

if you read carefully IPv6 RFC, you will notice that there is a separation between hosts and network infrastructure. What is specifically stated network infrastructure does not require global IPv6 addresses. All communication and routing can and should be done using LL addresses. That is the reason why pppoe-client does not add SLAAC address on the pppoe-clientt interface. On server side this feature is for hosts, that will connected and get simple SLAAC address that does not require DHCP-PD, like windows pppoe-client.

There is some other problem why dhcp-pd client does not get the prefix from your dhcp-pd server. And that should not be connected to fact that SLAAC address is not set on the router pppoe client interface.

Thu Dec 01, 2011 9:16 am

*) hotspot - fixed login page to better handle big load;
Can someone please elaborate a bit more on this fix?
As almost everyone has modified the original hotspot login page to his needs.
Cheers

There was a bug when there is a high load on hotspot service the new login pages didn't show up for the new clients - new version has a fix for it. This is doesn't effect your custom login pages, so you don't need to worry.

Eising · Thu Dec 01, 2011 12:00 pm

Impressive download speed indeed. Here from Denmark:

100%[========================================================================================================>] 19,199,504  15.7M/s   in 1.2s

angboontiong · Thu Dec 01, 2011 1:40 pm

It's the wirelss only improve on the 11N? what about the 11a?

and, what's the improvement if we select the rate to advance?

bambangs2komputer · Thu Dec 01, 2011 2:13 pm

Why upgrade-firmware in RB750 & 750G not same after upgrade v5.9 ?
RB750G

[bambang@VEKTOR ATAS JL. A. YANI NO5 TUREN Atas by ALDAMPITY] > system resource print 
                   uptime: 3m4s
                  version: 5.9
              free-memory: 17240KiB
             total-memory: 29708KiB
                      cpu: MIPS 24Kc V7.4
                cpu-count: 1
            cpu-frequency: 800MHz
                 cpu-load: 2%
           free-hdd-space: 33412KiB
          total-hdd-space: 61440KiB
  write-sect-since-reboot: 204
         write-sect-total: 1748
               bad-blocks: 0%
        architecture-name: mipsbe
               board-name: RB750G
                 platform: MikroTik
[bambang@VEKTOR ATAS JL. A. YANI NO5 TUREN Atas by ALDAMPITY] > system routerboard print 
       routerboard: yes
             model: 750G
     serial-number: 268D01734429
  current-firmware: 2.38
  upgrade-firmware: 2.38


RB 750
[bambang@VEKTOR.NET Bawah Jl. A.Yani No.5 Turen by ALDAMPITY] > system resource print 
                   uptime: 1h9m9s
                  version: 5.9
              free-memory: 15428KiB
             total-memory: 29708KiB
                      cpu: MIPS 24Kc V7.4
                cpu-count: 1
            cpu-frequency: 400MHz
                 cpu-load: 8%
           free-hdd-space: 28656KiB
          total-hdd-space: 61440KiB
  write-sect-since-reboot: 825
         write-sect-total: 6921087
               bad-blocks: 0%
        architecture-name: mipsbe
               board-name: RB750
                 platform: MikroTik
[bambang@VEKTOR.NET Bawah Jl. A.Yani No.5 Turen by ALDAMPITY] > system routerboard print 
       routerboard: yes
             model: 750
     serial-number: 28BC01A0F183
  current-firmware: 2.37
  upgrade-firmware: 2.37

Thu Dec 01, 2011 2:15 pm

each routerboard model has different RouterBOARD firmware. versions are not same for all RouterBOOT versions. this is normal.

paoloaga · Thu Dec 01, 2011 2:25 pm

From Italy, download.mikrotik.com has around 10mbit/s throughput and download2.mikrotik.com around 100mbit/s throughput.
download.mikrotik.com works with IPv6
download2.mikrotik.com is IPv4 only

my two cents.

aliyuim · Thu Dec 01, 2011 3:39 pm

hi, im new to this forum and i have a RB1100 and would want to have user manager. my Router Board is RB1100 OS v4.15.

Thu Dec 01, 2011 3:41 pm

hi, im new to this forum and i have a RB1100 and would want to have user manager. my Router Board is RB1100 OS v4.15.

post a new topic in the User Manager section, this is the wrong topic

angboontiong · Thu Dec 01, 2011 3:52 pm

hi, im new to this forum and i have a RB1100 and would want to have user manager. my Router Board is RB1100 OS v4.15.
post a new topic in the User Manager section, this is the wrong topic

Normis..
The Ethernet Port flapping still no solve with this firmware.

Thu Dec 01, 2011 3:57 pm

The Ethernet Port flapping still no solve with this firmware.

have you emailed support with supout.rif which was made at the moment of the issue? we still don't have any reliable reports.

angboontiong · Thu Dec 01, 2011 4:49 pm

The Ethernet Port flapping still no solve with this firmware.
have you emailed support with supout.rif which was made at the moment of the issue? we still don't have any reliable reports.

Dear Normis..
we have a lot of RB411 having this issue..

will grab all of it and email to support.

thanks.

Beeski · Thu Dec 01, 2011 5:06 pm

upgraded an RB800 with R52Hn cards that is an AP, running NV2 802.11n 5Ghz
wireless tx/rx signal strength improved 3 to 4 db

WirelessRudy · Thu Dec 01, 2011 6:37 pm

The Ethernet Port flapping still no solve with this firmware.
have you emailed support with supout.rif which was made at the moment of the issue? we still don't have any reliable reports.

Give me a break! What do you consider "reliable reports"??
You must have a ton of reports about this issue, and it is extensively discussed and mentioned in the forum in several related topics, some of these you even showed up!
I myself send several mails to you guys and even had some response to them from you guys. The main response has been; "we cannot reproduce the issue" and than some senseless suggestions are made! Like "give us access to your unit to look"! And than what, you are going to sit and wait hours, days to see it happening and than tell us you saw it?
I have a network with some 250 CPE's. All MT and all almost the same config. Only IP differs and name off course.
They are all always updated to the one but latest version and I already tried some with 5.9 too. I have set all units to e-mail alert if the Ethernet status changes and thus am building a nice overview on which units are effected and which ones not.
On average some 40% of my units are affected at times, of this some 5% might see it very regular and only a small number sees it happen several times everyday again. (And clients notice nothing!)
The issue is there since the v5.x family saw light and is since than also been discussed.

The problem of the whole issue is that is it not reproducible so you never know when it happen and how often. Some units have it several times some days and than work fine for days. Just when you think the problem is gone it just comes back!

I send one or two suppout.rif's of units that had it happen to you guys but obviously it didn't help. I made several reports and explanation about it on this forum too as most of you guys know as many other readers can acknowledge. So please don't act you are not aware of the problem!

My personal opinion is still it has something to do with the Ethernet ports, more in special the POE ones, that either became more sensible for ESD or strange frequencies (Ethernet traffic also travels with a frequency) or the sensor tricking the Ethernet status became more sensible for small deviations or errors. Maybe just a production change that causing this.

Something I recently discovered and could serve as a small fingerprinting is the following:
A client that has his contract suspended for half a year wanted the service enabled again.
After he connected the power back I hooked him up as wireless dhcp-client to my AP like normal and arranged some other stuff and off he went. Nice speed, good latency and no problems for him.
After two days setting new rules in my Outlook mail client to process the bulk of alert mails (port flaps) I'm getting each day, I noticed that this specific client produces more than 50 port flaps per hour! Still the client told me he had a good connection!
So we went to his place and found actually some slight green (copper) corrosion in the inside of the standard ('old' model ivory coloured triangle shaped with green led) powershot.
The guy told me it was an unheated room for years and indeed quit dampy in the winter.
We replaced the plug and the powershot (and sealed it this time) and since than we didn't see a single port flap any more......

In the last months I also had some main units (rb433's or 433AH's) that developed more serious port flaps. Here traffic really stopped to flow over the port and in one case it even made the board to reboot (NO watchdog enabled!) very regularly. In all these cases the issue almost completely disappeared by feeding the board by jack and at the same time not using the POE ports any longer. (These are weak anyway, I have a whole stack of boards with failed PoE ports..)

Several units with the issue show that although both sides are 100Mb able the port status is for days stable at 100Mb and than suddenly when the port flap happens the rate is set to 10Mb 9 out of 10 times. Sometimes it stays for days at 10Mb and only after the next 'port flap' storm it suddenly starts to use 100Mb again! Setting these units manually to 100Mb sometimes helps a bit, but not always or not completely.

One other finger pointing is that the issue on ALL units, so also my 24/7 running main routers, happens mainly during daytime hours. I would say that some 90% happen during ours the sun shines and the rest during the hours of darkness.
And that is not because usage is more during the day because in general my most heavy usage hours are in the evenings.
Only units that I have been able to supply with proper earthed connections where had the issue almost disappeared. Some of these units ran for years without a single problem! Only in ROS v5 it started.

One other thing I seem to notice is that the port flap usually never happens when there is really traffic running. Except these units where the ports are really disconnected because here the traffic also stops. For the many CPE with the issue this might explain that the client doesn't notice anything...

So I am more and more confident that it has something to do with ESD or just current leaking from the POE to the other connectors. That will than have an influence on the sensor that keep an eye on the status of the port. In some cases it might be even so bad it really starts to disconnect the port to negotiate a new communication rate. In some case I believe the issue is that severe it will take up to some seconds before a connection is really back up and running again...

It would be nice if MT finally starts working on this issue. I would bet much more users are having the issue. They only don't know because mostly clients don't notice and the ROS is not alerted on it by default. You have to set the proper logging options to actually being notified.

hajde · Thu Dec 01, 2011 8:35 pm

I think that your problem is hardware related. Try to replace the capacitors, on some RB-s help that.

On this problem, I am sure replace capacitors will help: "In all these cases the issue almost completely disappeared by feeding the board by jack and at the same time not using the POE ports any longer."

Djelibeybi · Thu Dec 01, 2011 11:02 pm

There is some other problem why dhcp-pd client does not get the prefix from your dhcp-pd server. And that should not be connected to fact that SLAAC address is not set on the router pppoe client interface.

Yup, so further research has shown me as well. It was just a guess.

I'm trying to grab packet sniffer reports from RouterOS and a working WIDE-DHCPv6-CLIENT install so that you guys can compare, but I'm not having much luck getting Packet Sniffer on RouterOS to actually capture useful packets.

Jeremy23 · Thu Dec 01, 2011 11:26 pm

I'm trying to grab packet sniffer reports from RouterOS and a working WIDE-DHCPv6-CLIENT install so that you guys can compare, but I'm not having much luck getting Packet Sniffer on RouterOS to actually capture useful packets.

You should be able to set up a mirror port on the Routerboard under the Switch section and run the capture on your PC.

Djelibeybi · Thu Dec 01, 2011 11:33 pm

You should be able to set up a mirror port on the Routerboard under the Switch section and run the capture on your PC.

Thanks, that's a great suggestion.

Djelibeybi · Fri Dec 02, 2011 1:11 am

You should be able to set up a mirror port on the Routerboard under the Switch section and run the capture on your PC.

Sadly, this doesn't work, as I don't see the packets traveling over the PPPoE link. Will have to try and get RouterOS's packet sniffer to work instead.

WirelessRudy · Fri Dec 02, 2011 1:30 am

I think that your problem is hardware related. Try to replace the capacitors, on some RB-s help that.

On this problem, I am sure replace capacitors will help: "In all these cases the issue almost completely disappeared by feeding the board by jack and at the same time not using the POE ports any longer."

Well, we are still talking 20 to 40% of my routerboards here! This means more than 100 units... And you are suggesting it might not even be the total solution. All units are in production so this would mean a mayor project that would take weeks so not months and would create a lot of disturbance amongst my clients.
The issue happens on all kind of boards. rb112's, rb133C3's, rb 411's, 433's, grooves, SXT's. Basically every model MT ever produced.
Age ranges also from the very first boards until the ones purchased just some weeks ago. If replacing the capacitors would be the issue I would be very much surprised not ALL MT users were affected. I have a very average sample of the total production line of MT over the years.

So thanks for the suggestion but I don't think this is the issue here. And if it is I'll guess I have to live with it and plan to move to another platform over time. The very last thing on my list!

Ascii · Fri Dec 02, 2011 4:54 am

I think that your problem is hardware related. Try to replace the capacitors, on some RB-s help that.

On this problem, I am sure replace capacitors will help: "In all these cases the issue almost completely disappeared by feeding the board by jack and at the same time not using the POE ports any longer."

I have the same problem with a RB750, inside a metallic cabinet, feeding by the original power supply (power jack), and connected to the switch with a 20cm cat5 cable.

15:05:12 interface,info ether1 link down
15:05:14 interface,info ether1 link up (speed 100M, full duplex)
15:08:16 interface,info ether1 link down
15:08:18 interface,info ether1 link up (speed 100M, full duplex)
15:11:12 interface,info ether1 link down
15:11:14 interface,info ether1 link up (speed 100M, full duplex)
15:15:04 interface,info ether1 link down
15:15:06 interface,info ether1 link up (speed 100M, full duplex)
16:01:08 interface,info ether1 link down
16:01:09 interface,info ether1 link up (speed 100M, full duplex)

The same switch connect to a RB433 without problem.
Port 2 of the RB750 connect to a Canopy AP without problems.

So I don´t thing thats is an ESD problem.

Another other one, RB411 ether1 to RB433 ether2
nov/24 11:27:57 interface,info ether1 link down
nov/24 11:27:59 interface,info ether1 link up (speed 100M, full duplex)
nov/25 10:25:55 interface,info ether1 link down
nov/25 10:25:57 interface,info ether1 link up (speed 100M, full duplex)
nov/25 10:50:12 interface,info ether1 link down
nov/25 10:50:14 interface,info ether1 link up (speed 100M, full duplex)
nov/25 11:19:15 interface,info ether1 link down
nov/25 11:19:18 interface,info ether1 link up (speed 100M, full duplex)
nov/25 15:30:11 interface,info ether1 link down
nov/25 15:30:14 interface,info ether1 link up (speed 100M, full duplex)
nov/26 20:05:31 interface,info ether1 link down
nov/26 20:05:33 interface,info ether1 link up (speed 100M, full duplex)

Always a 2 seconds port flapping, look fishing to me.

rodneal · Fri Dec 02, 2011 6:47 am

Hey All,

Just finished our network upgrade to v5.9 - happy to report ALL IS GREAT! From RB112s to the newest boards no lockups and no tower climbs - BH connections operating better - 2.4 APs seem slightly better in noisy environs.
Glad to have it!
Good Job Normis and MT Team!

Thank you,
Rod

Fri Dec 02, 2011 7:44 am

Hey All,

Just finished our network upgrade to v5.9 - happy to report ALL IS GREAT! From RB112s to the newest boards no lockups and no tower climbs - BH connections operating better - 2.4 APs seem slightly better in noisy environs.
Glad to have it!
Good Job Normis and MT Team!

Thank you,
Rod

No Problems with Rb112s? Tested 133c3 in lab and that board performed better with
5.6.

SXMU · Fri Dec 02, 2011 9:47 am

I think that your problem is hardware related. Try to replace the capacitors, on some RB-s help that.

On this problem, I am sure replace capacitors will help: "In all these cases the issue almost completely disappeared by feeding the board by jack and at the same time not using the POE ports any longer."
I have the same problem with a RB750, inside a metallic cabinet, feeding by the original power supply (power jack), and connected to the switch with a 20cm cat5 cable.

15:05:12 interface,info ether1 link down
15:05:14 interface,info ether1 link up (speed 100M, full duplex)
15:08:16 interface,info ether1 link down
15:08:18 interface,info ether1 link up (speed 100M, full duplex)
15:11:12 interface,info ether1 link down
15:11:14 interface,info ether1 link up (speed 100M, full duplex)
15:15:04 interface,info ether1 link down
15:15:06 interface,info ether1 link up (speed 100M, full duplex)
16:01:08 interface,info ether1 link down
16:01:09 interface,info ether1 link up (speed 100M, full duplex)

The same switch connect to a RB433 without problem.
Port 2 of the RB750 connect to a Canopy AP without problems.

So I don´t thing thats is an ESD problem.

Another other one, RB411 ether1 to RB433 ether2
nov/24 11:27:57 interface,info ether1 link down
nov/24 11:27:59 interface,info ether1 link up (speed 100M, full duplex)
nov/25 10:25:55 interface,info ether1 link down
nov/25 10:25:57 interface,info ether1 link up (speed 100M, full duplex)
nov/25 10:50:12 interface,info ether1 link down
nov/25 10:50:14 interface,info ether1 link up (speed 100M, full duplex)
nov/25 11:19:15 interface,info ether1 link down
nov/25 11:19:18 interface,info ether1 link up (speed 100M, full duplex)
nov/25 15:30:11 interface,info ether1 link down
nov/25 15:30:14 interface,info ether1 link up (speed 100M, full duplex)
nov/26 20:05:31 interface,info ether1 link down
nov/26 20:05:33 interface,info ether1 link up (speed 100M, full duplex)

Always a 2 seconds port flapping, look fishing to me.

At a history of ethernet flapping - my request to MT Team without answer :

"It's nice to see solid nv2 on ROS5.9 working, but all over the wireless pleasure and speeds i'm still having issues with ethernet interfaces, which stop
possibilities of nv2 to work with it. As same as topic on http://forum.mikrotik.com/viewtopic.php ... 50#p285372,

i have 1g ethernet interfaces freezing(actually stopping to transmit data) same on 5.5 5.6 5.7 5.8 5.9 of ROS, but have a solid ethernet connections on 4.17.

It is possible to insert on 4.17 wireless nv2 package same improvements as 5.8 or 5.9 packages of ROS 5.x series contains, as that was done for ROS
4.17 from 5.0-rc11? Seems to be perfect in my situation.

Attaching supout files <before> - supout.rif (normal state) and <after> - after-crash.rif (at this time ethernet was freezed, ind i accessed to router from wireless port)"

Thanks

...Something going bad with ether interfaces from ROS 5.x

Fri Dec 02, 2011 9:52 am

Actually, we don't! We don't have one customer who can tell us if the LED turns off or not, during the flapping. Please if anyone besides Wirelessrudy has seen any port flapping, please write to support. We don't have anyone else with a supout.rif in support.

Ibersystems · Fri Dec 02, 2011 10:22 am

- Locked up RB 750GL - unable to remote to the router.

- Lights up blinking on all ethernet ports.

- unable to reset

Ready to dismantle the Router and to be sent for RMA
this is definitely not related to v5.9 or any other RouterOS version. you got yourself a hardware issue, contact support

My 5 brand new 411AR looked up when I upgraded from 5.8 to 5.9. Leds off. Power cycle rebooted ok. Anyone had problems like this with 433AH or 433?

thanks,

hajde · Fri Dec 02, 2011 11:57 am

Same problem with x86. After upgrade on 5.9 its locked after reboot everything ok.

hajde · Fri Dec 02, 2011 12:13 pm

I think that your problem is hardware related. Try to replace the capacitors, on some RB-s help that.

On this problem, I am sure replace capacitors will help: "In all these cases the issue almost completely disappeared by feeding the board by jack and at the same time not using the POE ports any longer."
Well, we are still talking 20 to 40% of my routerboards here! This means more than 100 units... And you are suggesting it might not even be the total solution. All units are in production so this would mean a mayor project that would take weeks so not months and would create a lot of disturbance amongst my clients.
The issue happens on all kind of boards. rb112's, rb133C3's, rb 411's, 433's, grooves, SXT's. Basically every model MT ever produced.
Age ranges also from the very first boards until the ones purchased just some weeks ago. If replacing the capacitors would be the issue I would be very much surprised not ALL MT users were affected. I have a very average sample of the total production line of MT over the years.

So thanks for the suggestion but I don't think this is the issue here. And if it is I'll guess I have to live with it and plan to move to another platform over time. The very last thing on my list!

You have hardware which is 3 or more years old, it is normal that "my grandmother" is dead. Especially because she live all time outside.

I just suggest what has helped on SOME of my devices. Just try it, especially on old RB. You must have some of them which is not on productions. I am 100% sure that this problem when is everything ok on jack and not on POE, that is capacitors problem. Like I said: Just try it.

emersonmill · Fri Dec 02, 2011 1:10 pm

Hello everyone.
My RB1200 presents the problem of ethernet port flaps, the LED turns off.

On the reboot cycles is probably the problem with the firewall, the firewall was sensitive 5.x versions.

The order of rules has to be perfect to not overload the system.

The layer7 has problems if you use the PPPOE connections only layer7 not overload the system.

If you use the connections routed layer7 cause overload in the system.

DNS problems with this slow.

The routing table with this problem of slow.

All these problems are being generated by an error in the routing table.

The MK team is solving the easy problems.

We're the guinea pigs !!!!

KillerOPS · Fri Dec 02, 2011 1:54 pm

download faster than before.
Before i got like 500KB/s, now it gets pretty easily to 2MB/s (bytes)
Location is Romania through euroweb-telia

Thanks

angboontiong · Fri Dec 02, 2011 2:45 pm

I think that your problem is hardware related. Try to replace the capacitors, on some RB-s help that.

On this problem, I am sure replace capacitors will help: "In all these cases the issue almost completely disappeared by feeding the board by jack and at the same time not using the POE ports any longer."
Well, we are still talking 20 to 40% of my routerboards here! This means more than 100 units... And you are suggesting it might not even be the total solution. All units are in production so this would mean a mayor project that would take weeks so not months and would create a lot of disturbance amongst my clients.
The issue happens on all kind of boards. rb112's, rb133C3's, rb 411's, 433's, grooves, SXT's. Basically every model MT ever produced.
Age ranges also from the very first boards until the ones purchased just some weeks ago. If replacing the capacitors would be the issue I would be very much surprised not ALL MT users were affected. I have a very average sample of the total production line of MT over the years.

So thanks for the suggestion but I don't think this is the issue here. And if it is I'll guess I have to live with it and plan to move to another platform over time. The very last thing on my list!

You have hardware which is 3 or more years old, it is normal that "my grandmother" is dead. Especially because she live all time outside.

I just suggest what has helped on SOME of my devices. Just try it, especially on old RB. You must have some of them which is not on productions. I am 100% sure that this problem when is everything ok on jack and not on POE, that is capacitors problem. Like I said: Just try it.

we have tried using the DC jack instead of POE.
the problem still remain, btw, which capacitor you talking about.

can post the picture here and we may can have a try on it.
if it work, then why not mikrotik change the item when it on the production?

sound non sense.

Lakis · Fri Dec 02, 2011 3:49 pm

Same problem with x86. After upgrade on 5.9 its locked after reboot everything ok.

with 5.9 my x86 reboot evry 5-6 min
And still my rb1200 reboot 4-5 times a day

timberwolf · Fri Dec 02, 2011 7:23 pm

And if it is I'll guess I have to live with it and plan to move to another platform over time. The very last thing on my list!

As if there were any comparable alternative...

hajde · Fri Dec 02, 2011 11:02 pm

[/quote]

we have tried using the DC jack instead of POE.
the problem still remain, btw, which capacitor you talking about.

can post the picture here and we may can have a try on it.
if it work, then why not mikrotik change the item when it on the production?

sound non sense.[/quote]

I think that we do not understand each other.

When I say that the routers is in production, I think that routers is somewhere installed, for example at client side...

When you see the condenser which is inflated / puffed, from time consuming and labor, it means that it is incorrect and should be replaced.

Broken / old capacitor can cause such errors to the router restarts. In most cases, if is failure by the capacitor, the use of Jack instead of POE temporarily solve the problem.

In version 5.x I also had problems with the restart. Reverting to version 4.17 I solved the problem. While this problem with jack connector and POE I have resolved with changing the capacitors.

hajde · Fri Dec 02, 2011 11:04 pm

Same problem with x86. After upgrade on 5.9 its locked after reboot everything ok.
with 5.9 my x86 reboot evry 5-6 min
And still my rb1200 reboot 4-5 times a day

Your x86 is Multi core (core2duo or similar procesor)?

rodneal · Fri Dec 02, 2011 11:49 pm

Hey All,

Just finished our network upgrade to v5.9 - happy to report ALL IS GREAT! From RB112s to the newest boards no lockups and no tower climbs - BH connections operating better - 2.4 APs seem slightly better in noisy environs.
Glad to have it!
Good Job Normis and MT Team!

Thank you,
Rod
No Problems with Rb112s? Tested 133c3 in lab and that board performed better with
5.6.

No - We have v5.9 loaded on a few old WRAP boards and PC systems (x86) and RB112 and RB133 and RB532's - the RF seems to perform better. No crashes. All good.
Rod

angboontiong · Sat Dec 03, 2011 12:03 pm

Hey All,

Just finished our network upgrade to v5.9 - happy to report ALL IS GREAT! From RB112s to the newest boards no lockups and no tower climbs - BH connections operating better - 2.4 APs seem slightly better in noisy environs.
Glad to have it!
Good Job Normis and MT Team!

Thank you,
Rod
No Problems with Rb112s? Tested 133c3 in lab and that board performed better with
5.6.
No - We have v5.9 loaded on a few old WRAP boards and PC systems (x86) and RB112 and RB133 and RB532's - the RF seems to perform better. No crashes. All good.
Rod

Dear Mikrotik..
can be the wireless to make like the microwave?

which can carry more bandwidth..

angboontiong · Sat Dec 03, 2011 12:04 pm

we have tried using the DC jack instead of POE.
the problem still remain, btw, which capacitor you talking about.

can post the picture here and we may can have a try on it.
if it work, then why not mikrotik change the item when it on the production?

sound non sense.[/quote]

I think that we do not understand each other.

When I say that the routers is in production, I think that routers is somewhere installed, for example at client side...

When you see the condenser which is inflated / puffed, from time consuming and labor, it means that it is incorrect and should be replaced.

Broken / old capacitor can cause such errors to the router restarts. In most cases, if is failure by the capacitor, the use of Jack instead of POE temporarily solve the problem.

In version 5.x I also had problems with the restart. Reverting to version 4.17 I solved the problem. While this problem with jack connector and POE I have resolved with changing the capacitors.[/quote]

any symptom or measurement for us to make sure that's the capacitor issue?
appreciate it..

thanks.

hajde · Sat Dec 03, 2011 12:27 pm

any symptom or measurement for us to make sure that's the capacitor issue?
appreciate it..

thanks.[/quote]

I think that you have not problem with capacitors. Try v.4.17 and you will have bether dreams.

WirelessRudy · Sat Dec 03, 2011 2:46 pm

I think that your problem is hardware related. Try to replace the capacitors, on some RB-s help that.

On this problem, I am sure replace capacitors will help: "In all these cases the issue almost completely disappeared by feeding the board by jack and at the same time not using the POE ports any longer."
Well, we are still talking 20 to 40% of my routerboards here! This means more than 100 units... And you are suggesting it might not even be the total solution. All units are in production so this would mean a mayor project that would take weeks so not months and would create a lot of disturbance amongst my clients.
The issue happens on all kind of boards. rb112's, rb133C3's, rb 411's, 433's, grooves, SXT's. Basically every model MT ever produced.
Age ranges also from the very first boards until the ones purchased just some weeks ago. If replacing the capacitors would be the issue I would be very much surprised not ALL MT users were affected. I have a very average sample of the total production line of MT over the years.

So thanks for the suggestion but I don't think this is the issue here. And if it is I'll guess I have to live with it and plan to move to another platform over time. The very last thing on my list!

You have hardware which is 3 or more years old, it is normal that "my grandmother" is dead. Especially because she live all time outside.

I just suggest what has helped on SOME of my devices. Just try it, especially on old RB. You must have some of them which is not on productions. I am 100% sure that this problem when is everything ok on jack and not on POE, that is capacitors problem. Like I said: Just try it.

You are not reading well my friend.. As you can see the issue exists on SXT and Groove's as well. Apart from that they are not even one year in production there is no way to repair or replace parts on these boards without distroying them...... Hence your soluction is nice but not workable.
Secondly, these units all work fine in 3.x and 4.x version.
Some more info to consume; some units having the issue don't have the issue when fit at another location. All that changes is the cable, maybe the adapter, sometimes the powershot and the RJ-45.
It also works the other way round. Some units never had the issue but suddenly at different location they have. Put it at yet again different location and the problem is gone again....

This all cannot happen because of some 'leaking' or melted capacitors.....

WirelessRudy · Sat Dec 03, 2011 2:59 pm

any symptom or measurement for us to make sure that's the capacitor issue?
appreciate it..

thanks.

I think that you have not problem with capacitors. Try v.4.17 and you will have bether dreams.

Keep dreaming. Because the 4.17 can't log the Ethernet port status it doesn't mean the issue is gone. You just don't see it ....... So it is a matter of guessing if it is existing in 4.17 or not.
Since 5.x has several improvements that 4.x doesn't have I want to use 5.9. If there is an issue it has to be solved.

Because your car is not performing well do you go back to the dealer to order horse and coach?
We are not living in the past. New software/hardware always come with issues/bugs and the only way to improve and move forward is to find a solution to these. Going back in time is not one of them....

The only problem here is that the issue is reported by several users but all the rest including MT can't see or reproduce it so they think it doesn't exist. I'll guess we users beeing effected by it have to find out ourselves.....

hajde · Sat Dec 03, 2011 4:24 pm

You are not reading well my friend.. As you can see the issue exists on SXT and Groove's as well. Apart from that they are not even one year in production there is no way to repair or replace parts on these boards without distroying them...... Hence your soluction is nice but not workable.
Secondly, these units all work fine in 3.x and 4.x version.
Some more info to consume; some units having the issue don't have the issue when fit at another location. All that changes is the cable, maybe the adapter, sometimes the powershot and the RJ-45.
It also works the other way round. Some units never had the issue but suddenly at different location they have. Put it at yet again different location and the problem is gone again....

This all cannot happen because of some 'leaking' or melted capacitors.....

Yes, there are different situations, and this is normal. I think that I exchange opinions, with a man who works at an ISP and to whom is normal that there are different situations.

I know you're frustrated with the constant problems with clients and reboot the router.

But I do not work in Mikrotik, I am just trying to help you.

mramos · Sun Dec 04, 2011 12:55 am

Hi ...

Some notes after upgrade from 5.8 to 5.9 pre release (Nov 14th) => 5.9 final - AND - 2.38 fw where it applies.

Boards are RB433 (2), 433AH (1), 433UAH (1), WRAP (2, Geode X86), 411AR (3, ufl & MMCX). Cards are R52Hn, R52nM, R52n, XR9. Setup are PMP on 2.4GHz, 20 & 10MHz channels (Only G), UBNT CPEs, max 30 authenticated CPEs each radio card, total 130 CPEs.

- Weak signal performance has improved. Some CPEs still at the -78 ... -82 range but now keep connected for days. Easy access to CPE gui (used to be hard sometimes, despite of pings < 5ms).

- PMP at 900MHz, Nstreme, 5MHz bw, NLOS from AP to CPEs: better troughtput and - better latency when loaded. Works ok on NV2 (even with a WRAP X86 at one side), more tests required due to VOIP issue (ping times ~ 2 ... 5ms but voice quality sucks, the same does not occurr when Nstreme is used). Noise measurement better than older versions (used to jump suddenly from nominal -92dBm to -80 something and link drops). Distance on NV2 on previous versions swings from 5 to 15Km on a 2.4Km NLOS. Now steady 5Km. Nice.

- On 20MHz bw (R52nM, max 30 authenticated CPEs), if I set distance manually (2Km), use advanced rate selection, manually configured AND left noise reduction enabled, after some time CPEs stop answering to pings (start rising from <5ms then 400ms then timeout). As soon as I use distance = dynamic, rate selection = legacy, noise reduction = off everything ok for days.

- On 10MHz bw (R52Hn, max 20 authenticated CPEs) I can keep distance = 5Km (real max = 2Km) but if left rate selection = advanced, data rates = configured, noise reduction = enabled after some time CPEs start disconecting and connecting back. So it's configured the same way as 20MHz bw is, except in this case I can keep distance = manually configured.

In general: all smooth, no more reboots, reliable.

P.S. I spent (or invest) some days exchanging e-mails with Janis about an issue with router random reboots (Kernel failure) and finally the source was discovered: UBNT Nano2 LocoM (running V 5.3.3) have an issue with WPA2-PSK (AES) which - depending on how many of those model are authenticated - cause ROS Kernel Failures.

I changed all security setup to WPA-PSK (TKIP) and the problem gone. More sensitive on 20MHz channel than in 10MHz channel.

The same crashes hapened in two different places and vanish as soon as I changed all N2LM to WPA-PSK (later locked to WPA-PSK TKIP).

This issue was around since june/july when I started to mix Bullet2/Nano2 Loco with Nano2 LocoM.

Regards;

Lakis · Mon Dec 05, 2011 3:18 am

Same problem with x86. After upgrade on 5.9 its locked after reboot everything ok.
with 5.9 my x86 reboot evry 5-6 min
And still my rb1200 reboot 4-5 times a day

Your x86 is Multi core (core2duo or similar procesor)?

My x86 is http://www.mikrotikrouter.com/732.asp

on RB1200 I turn off Conn. Tracking because for now I use him like a switch but still reboots 1 or 2 times a day

Mon Dec 05, 2011 1:28 pm

For RB1200 rebooting you need to write to support and get pre-production v5.10 with the fix.

For all port flappers - each Ethernet link have 2 ends and a wire in the middle, so 3 potential points of failure or simple incompatibility between 2 devices.
AFAIK port linking is a hardware feature and as WirelessRuby explained, only thing new in software was log messages.

I have seen port flapping for many different reasons:
1) old Cisco devices didn't talk normally to other devices
2) not shielded Ethernet cable near high voltage power lines
3) shielded, but not grounded Ethernet cables
4) Partially blown Ethernets on boards, because of lack of grounding
5) some POE injector problems
6) too long or too short Ethernet cable

So from where i stand - all those "I upgraded and my ports started to flap" problems makes me smile. "yea, right!"

Mon Dec 05, 2011 1:34 pm

I can agree to some of your points. To clarify again to others - we only recently added the ethernet status to log. Your ports probably always "flapped" but only now you see it in the log.

honzam · Mon Dec 05, 2011 4:14 pm

Your ports probably always "flapped" but only now you see it in the log.

+1. I also think. Mostly it is a faulty cable or connector

THG · Mon Dec 05, 2011 5:15 pm

can you all please test the "Latvia" download mirror? What speeds to you get in your country (name it). Is it higher than ususal?

Country: Sweden

Download: mikrotik-5.9.iso

Time:

http://download.mikrotik.com = 1 minute
http://download2.mikrotik.com = 1½ seconds

Slightly faster than before.

hajde · Mon Dec 05, 2011 5:22 pm

My x86 is http://www.mikrotikrouter.com/732.asp

[/quote]

I know that this is not the final solution that I shall offer, but may reduce pain and problems with clients until the final solution from Mikrotik.
This is software problem becouse I also have 3 MT on x86 with same simptoms.

Problem is with multi-core, just off the support for multi-core and you will have not longer any restart.

changeip · Mon Dec 05, 2011 6:35 pm

I can agree to some of your points. To clarify again to others - we only recently added the ethernet status to log. Your ports probably always "flapped" but only now you see it in the log.

I agree this probably isnt really affecting traffic, however, is it more a case of the software reporting flaps when there really isnt any? Maybe it's opposite, you think it was previously flapping, but I think maybe it has been solid and the flaps are just misreported (because of a missed poll or something).

omega-00 · Tue Dec 06, 2011 6:40 am

From: http://wiki.mikrotik.com/wiki/Manual:IPv6_PD_over_PPP

/ppp profile set default dhcpv6-pd-pool=myPool

Is this section required to be configured on the client device? In the same way?

ranjan · Tue Dec 06, 2011 6:42 am

do you need more official than the official webpage?
http://www.mikrotik.com/download.html

Hello Normis,
I am using MT- RB450G router board and it has OS v5.6 with level 5 license.
Can I upgrade this to v5.9 without disturbing the license?
Do I will have to re-configure the router borad again for my usage or upgrade will not disturb my configuration?
Please look at the attached image and tell me should I go for upgradation or not?

Thanks in Advance.

omega-00 · Tue Dec 06, 2011 6:45 am

Hi Ranjan,

You can upgrade all the way to anything in v6 with your current licence. We are only up to version v5.9 at the moment so you are still fine to continue upgrading.
The upgrade should not interfere with your config in any way however it is always good practice to test somewhere you won't break anything first

Hope this helps!

Djelibeybi · Tue Dec 06, 2011 6:47 am

Is this section required to be configured on the client device? In the same way?

No, that's for the server side. It's under the big Server heading and everything.

ranjan · Tue Dec 06, 2011 7:59 am

Hi Ranjan,

You can upgrade all the way to anything in v6 with your current licence. We are only up to version v5.9 at the moment so you are still fine to continue upgrading.
The upgrade should not interfere with your config in any way however it is always good practice to test somewhere you won't break anything first

Hope this helps!

Thanks omega-00
I will try to upgrade today only.

omega-00 · Tue Dec 06, 2011 9:24 am

No, that's for the server side. It's under the big Server heading and everything.

I am aware of the article and what it says, however given it says "PPP will use specified IPv6 pool to create dynamic DHCP server." I was curious as to if it was also relevant / able to be used to define the pool that will be allocated to the client's LAN side (ethernet attached) DHCP6 server(s).

The PPP listing for this says much the same thing.

http://wiki.mikrotik.com/wiki/Manual:PP ... r_Profiles

dhcpv6-pd-pool (string; Default: ) Name of the IPv6 pool which will be used by dynamically created DHCPv6-PD server when client connects.

Asking questions about a new feature is the only way I can find out, given the problem we're having getting our links online for real could just as easily be on Internode's side as the MikroTik's

Djelibeybi · Tue Dec 06, 2011 10:33 am

http://wiki.mikrotik.com/wiki/Manual:PP ... r_Profiles
dhcpv6-pd-pool (string; Default: ) Name of the IPv6 pool which will be used by dynamically created DHCPv6-PD server when client connects.

But we're not running a DHCPv6-PD server, just a client. So we don't need to configure a DHCPv6-PD prefix pool. We get given one by Internode, or at least, we should be given one by Internode.

omega-00 · Tue Dec 06, 2011 2:08 pm

But we're not running a DHCPv6-PD server, just a client. So we don't need to configure a DHCPv6-PD prefix pool. We get given one by Internode, or at least, we should be given one by Internode.

So how am I supposed to pass prefixes to other routers on my network?

Djelibeybi · Tue Dec 06, 2011 10:25 pm

So how am I supposed to pass prefixes to other routers on my network?

This is done by RADVD not by DHCPv6-PD. Mikrotik have found the problem and provided me with a test version of RouterOS, which I'm about to download. Will report back.

WirelessRudy · Wed Dec 07, 2011 1:34 am

For RB1200 rebooting you need to write to support and get pre-production v5.10 with the fix.

For all port flappers - each Ethernet link have 2 ends and a wire in the middle, so 3 potential points of failure or simple incompatibility between 2 devices.
AFAIK port linking is a hardware feature and as WirelessRuby explained, only thing new in software was log messages.

I have seen port flapping for many different reasons:
1) old Cisco devices didn't talk normally to other devices
2) not shielded Ethernet cable near high voltage power lines
3) shielded, but not grounded Ethernet cables
4) Partially blown Ethernets on boards, because of lack of grounding
5) some POE injector problems
6) too long or too short Ethernet cable

So from where i stand - all those "I upgraded and my ports started to flap" problems makes me smile. "yea, right!"

6) too long or too "short" Ethernet cable"? Too short? Since when can a cable be too short? This is new to me...

bradg · Wed Dec 07, 2011 2:12 am

Too short? Since when can a cable be too short? This is new to me...

Actually, the National Semiconductor DP83815 and DP838156 Ethernet silicon had hardware related issues with short cables. There is a setting specifically for those cards in ROS - see http://wiki.mikrotik.com/wiki/Manual:Interface/Ethernet

I believe those Ethernet IC's were used in all WRAP series (not ALIX, which uses VIA Ethernet silicon) boards, as well as all Soekris net4XXX series units. It could also be used on the RB2XX boards, but I'm not sure as I don't have any in my network.

So, yes, it is possible to have too short of Ethernet cable

--Brad

omega-00 · Wed Dec 07, 2011 3:01 am

Edit: brad beat me to it but yeah, there is too short a cable.

http://en.wikipedia.org/wiki/Ethernet_p ... le_lengths

It's a case of "most people don't know until they manage to break it", there even used to be a cable length setting in v3 for specific ethernet cards. The more you know!

fewi · Wed Dec 07, 2011 4:32 am

Hopefully no one is using half duplex wired Ethernet connections anymore, though.

ranjan · Wed Dec 07, 2011 6:15 am

Hi Ranjan,

You can upgrade all the way to anything in v6 with your current licence. We are only up to version v5.9 at the moment so you are still fine to continue upgrading.
The upgrade should not interfere with your config in any way however it is always good practice to test somewhere you won't break anything first

Hope this helps!

Hi omega-00,
After upgrade if anything goes wrong and I want to rollback to v5.6 with all previous configuration then how can I do that?
I have taken backup (Files>Backup).
Can u please tell me step by step process for taking MT RB450G router to its previous version (v5.6) and making it to working condition?
Thanks in advance.

omega-00 · Wed Dec 07, 2011 6:22 am

drop v5.6 npk file in the root directory, head to system -> packages -> click downgrade.

ranjan · Wed Dec 07, 2011 6:43 am

drop v5.6 npk file in the root directory, head to system -> packages -> click downgrade.

Where to get all v5.6 package files.
Can u please provide me with the links? I am unable to get the package files on MT official website.

omega-00 · Wed Dec 07, 2011 6:48 am

Where to get all v5.6 package files.
Can u please provide me with the links? I am unable to get the package files on MT official website.

http://www.mikrotik.com/download/router ... .6.torrent

If you have any more questions on this topic please post a new thread as this is not relevant to the current one.

synnest · Wed Dec 07, 2011 7:47 pm

I'm using a rb411u with 3G modem Huawei em770w , the last working ros for me is 5.7, in 5.8 and 5.9 the modem simply won't connect.
Anybody else having issues with 3g modems in these latest versions?

jonashi · Wed Dec 07, 2011 9:50 pm

I'm having hard time using ROS5.9. I've upgraded from 4.10 and 4.17 (4x worked fine). After upgrade I got random reboots of both my core routers, both are x86 multicpu (xeon), one is 1xquadcore system http://h18000.www1.hp.com/products/quic ... 67_na.HTML and second is 2x quad core xenon system.

- in log I can see "router was rebooted without proper shutdown", randomly, sometimes two times a day, so I don't know hot to get suppout.rif from them when a situation occures.

Both roouters are simple, as their task are only simple queue with amount <800 and NATting. No L7 or routing BGP, OSPF.. or such. Just a simple setups. On both is set /system hardware multi-cpu =yes.

My other x86 machines w/o simple queue and NAT working fine with ROS5.9, even with more than 2000 queue tree queues, the same routerboards.

Did anyone other encountered the same strange difficulties on x86 systems ?

siprox · Wed Dec 07, 2011 11:15 pm

Yes, i have the same problems to, my x86 crash randomly after upgrade from 5.8 to 5.9.
my setup not to complex, just a couple of queue tree and webproxy internal.

mwheeler · Thu Dec 08, 2011 8:08 am

So how am I supposed to pass prefixes to other routers on my network?
This is done by RADVD not by DHCPv6-PD. Mikrotik have found the problem and provided me with a test version of RouterOS, which I'm about to download. Will report back.

Am I able to test out this version as well? Also an Internode Customer.

janisk · Thu Dec 08, 2011 8:54 am

From: http://wiki.mikrotik.com/wiki/Manual:IPv6_PD_over_PPP
/ppp profile set default dhcpv6-pd-pool=myPool
Is this section required to be configured on the client device? In the same way?

on client side you set up only in DHCP-client what the name of dynamic pool is going to be. Profile only has to be set on the server side to enable dynamic dhcp-server creation on the pppoe server interface.

Thu Dec 08, 2011 10:31 am

We have confirmed that the Internode issue has been fixed in v5.10, we will release it later today.

hajde · Thu Dec 08, 2011 10:39 am

I'm having hard time using ROS5.9. I've upgraded from 4.10 and 4.17 (4x worked fine). After upgrade I got random reboots of both my core routers, both are x86 multicpu (xeon), one is 1xquadcore system http://h18000.www1.hp.com/products/quic ... 67_na.HTML and second is 2x quad core xenon system.

- in log I can see "router was rebooted without proper shutdown", randomly, sometimes two times a day, so I don't know hot to get suppout.rif from them when a situation occures.

Both roouters are simple, as their task are only simple queue with amount <800 and NATting. No L7 or routing BGP, OSPF.. or such. Just a simple setups. On both is set /system hardware multi-cpu =yes.

My other x86 machines w/o simple queue and NAT working fine with ROS5.9, even with more than 2000 queue tree queues, the same routerboards.

Did anyone other encountered the same strange difficulties on x86 systems ?

Yes, same problem. Or multi-cpu=no or v4.x Than work without problem.

Problem is with simple queue, but I am not sure that mikrotik know for that problem.

jonashi · Thu Dec 08, 2011 10:57 am

@hajde
thanks
I consider it as confirmed on my side. I moved my current setup to DELL1850 -another machine, clean fresh instalation 5.9, then imported rsc step by step (firewall..queue...etc) problem still remains. I've downgraded to 4.17, unit is working fine now.

Thu Dec 08, 2011 11:29 am

jonashi, hajde

connect a keyboard and monitor, and try to photograph the Kernel panic that shows on the screen before it reboots. We need to see this. Send it to support as a photo.

first set /system console screen to 40 lines (to see all) and then use a high speed photo capture, or HD video at 60fps

Thu Dec 08, 2011 11:51 am

Same problem with x86. After upgrade on 5.9 its locked after reboot everything ok.
with 5.9 my x86 reboot evry 5-6 min
And still my rb1200 reboot 4-5 times a day

We have fixed this in the latest version v5.10, support can give it to you

angboontiong · Thu Dec 08, 2011 12:19 pm

Hi Normunds...

The V5.9 with the wds setting seem can provide more higher throughput...
due to the pc limitation, i only can do 1 way file transferring.

but 1 thinking i would highlight again, the port flapping issue still there.
if this can be solved, then it will be wonderful..

i have tested many site, the Ethernet running on 10BaseT mode will having no issue, but when put it on the auto mode, then the port flapping will occur.

is possible this is due to the Ethernet chipset issue?

thanks.

Thu Dec 08, 2011 12:40 pm

please read the post above. the issue is not in software. check cables and other end ports. make sure your ethernet has no hardware problem.

kyob · Thu Dec 08, 2011 2:33 pm

can you all please test the "Latvia" download mirror? What speeds to you get in your country (name it). Is it higher than ususal?

2011-12-08 13:32:12 (9,08 MB/s) - `mikrotik-5.9.iso' from Poland.

angboontiong · Thu Dec 08, 2011 4:40 pm

please read the post above. the issue is not in software. check cables and other end ports. make sure your ethernet has no hardware problem.

Dear Normunds..
yes, the cable and connector and switches have no issue at all.

when we replace with another one unit, there is no problem.
we experience in a lot of place.

thanks

mramos · Thu Dec 08, 2011 10:32 pm

Hi ...

After upgrade a RB433 to 5.9 I noticed that 3 laptops that are tied to an Engenius EOC1650 configured as AP-Bridge sometimes time out DHCP leases (IPV4) trhu 433 ETH.

Explaining better: the EOC1650 is connected to RB433 ETH port. There's a DHCP server on 433 which provide addresses to those laptops trhu ETH => EOC1650 => laptops.

I remember that this occurred firt time as soon as I upgraded from 5.7 to 5.8. Then I downgraded back to 5.7 to "solve" this problem.

Now I jumped from 5.7 to 5.9 and this behaviour came back.

At DHCP tables the message is (was) the same ... "offered ..." and that's it.

Regards;

goranst · Thu Dec 08, 2011 11:35 pm

Same problem with x86. After upgrade on 5.9 its locked after reboot everything ok.
with 5.9 my x86 reboot evry 5-6 min
And still my rb1200 reboot 4-5 times a day
We have fixed this in the latest version v5.10, support can give it to you

Hi!
Same problem on my x86 multi core HP server, but for me a bigger problem now is that the few links I have less traffic. The problem is that the link does not pull enough traffic as needed, when back-links in 5GHz-A Nstreme traffic grows rapidly than 5GHz-N Nstreme or NV2. Configurations are more or less standard classical PTP link. Is there any advice .. thank you.

Goran.

angboontiong · Fri Dec 09, 2011 4:17 am

We have fixed this in the latest version v5.10, support can give it to you[/quote]

Hi!
Same problem on my x86 multi core HP server, but for me a bigger problem now is that the few links I have less traffic. The problem is that the link does not pull enough traffic as needed, when back-links in 5GHz-A Nstreme traffic grows rapidly than 5GHz-N Nstreme or NV2. Configurations are more or less standard classical PTP link. Is there any advice .. thank you.

Goran.[/quote]

you can see that your signal is not good enough, to run the N, you signal should min <-55dbm with ccq is on 90%/90% and above.

omidkosari · Fri Dec 09, 2011 9:38 am

any progress in 100% cpu usage problem when monitorin router with dude ? http://forum.mikrotik.com/viewtopic.php?f=1&t=47655

goranst · Fri Dec 09, 2011 10:22 am

Thanks!

Hm, this is the link of 16km, what is your recommendation for more bandwidth, better signal is difficult to make. Should I go back to the 5GHz A-turbo that I drew more traffic, 5GHz-N was my option, but get -55dbm can be difficult to do. I still have a couple of links that are hungry for higher throughput, but also with the signal between -60dbm to -75dbm. What to do ...help!

Goran.

janisk · Fri Dec 09, 2011 10:25 am

go for higher gain antenna, if regulatory domain allows it

goranst · Fri Dec 09, 2011 12:21 pm

go for higher gain antenna, if regulatory domain allows it

I already have, I think a good antenna 35dBi Dualpol 900mm/AL Parabola, fresnel zone clear and still can not get a better signal. What is the minimum signal strength that the link worked well with the standard 5GHz-N with NV2 or Nstreme. Maybe I'm wrong basically, in the setting of new links N link. With v5.9 I got better performance, but on shorter links, long links are still a problem to me ...

In any case, thanks for the tip, at least I will know to look for the error.

Goran

janisk · Fri Dec 09, 2011 12:23 pm

in that case, check everything and put your link in calculator to see what power you should get for this link with antennas like that.

vaizki · Fri Dec 09, 2011 1:04 pm

We've been running 5.9 for 8 days now and the entire time the CPU load has been near 50%..

In /tool profile it's the "routing" process that's taking up CPU:

NAME CPU USAGE
ethernet all 0%
console all 0%
firewall all 0%
management all 0.5%
idle all 37.5%
queuing all 0.5%
routing all 45.5%
unclassified all 16%

We're running BGP to 2 peers on this box and before the upgrade we were at about 5% cpu usage. Anyone else run into this problem and a solution?

Edit: This is a RB1100.

Fri Dec 09, 2011 1:07 pm

Please send a supout file to support.

shyrwall · Wed Dec 14, 2011 8:37 am

Not sure if I should post this here but..

When flashing the BIOS of my RB450G to 2.38 the CPU Mhz gets set to 100Mhz.

Took a while to figure out because I upgraded to 5.9 at the same time.

Just a heads up if some people are complaining about performance problems after upgrading RouterOS.

otgooneo · Wed Dec 14, 2011 10:37 am

Does any one has X86 with ROS V5.9? My console, terminal, script, scheduler, telnet, netwatch, ssh don`t work. If any had X86 please check for me. Also what do I need to do? Previous was 5.0rc11 and worked fine.

NetworkPro · Wed Dec 14, 2011 3:16 pm

Reinstall that router from CD or NetInstall and maybe test for hardware issues.

otgooneo · Tue Dec 20, 2011 2:17 am

Today I have downgraded version to V5.5. Now all terminal, scheduler, scripts are working OK.