The device is an industrial grade rackmount. Rackmounts are usually meant to be installed in racks, in server rooms. We did not aim for low noise, we tried to improve temperature and performance in this case.

As for the reboot - if your device was continuously operating for two weeks, it means you don't have the latest versions. Please upgrade to the current RC7

Deleted because not related.

Did anybody use CCR in production environment ?
I'm trying to do some firewalling on ccr software v7, and i'm dissapointed with performance.
Same situation on 2 different environments :
1. Heavy 150 mbps udp traffic/ 80k pps. Losses from time to time, but this customer recieve a lot of ddoses and every small udp ddos ( 300 mbps ) causes ccr hang (looks like it stop to forward any traffic) until end of attack. Full firewall for input and forward chain, from non standard rules rate limiting for udp to 30 pkt / min.

2. Traffic around 300 mbps / 180 k pps, just rules for input to have access to ccr, same situation, CPU indicating 0 & but from time to time you cannot access to ccr and it not forward any traffic for the while, and it come back. Nothing in logs etc.

Best Regards
Michal

RC7 currently Firewall doesn't have optimization for 36 cores, we will deliver that as soon as possible, and your performance should increase some 25-30 times.

I still have a problem with Queue Tree.

As soon as i set the "parent" of the top queue to "global" - increases delays to hosts that describes by this rule. If i set as "parent" the interface - everything works fine.
For example - there is running a torrent download, computer`s address is added to "Mangle" and created a rule in the "Queue Tree" with a maximum speed at 100Mbit/sek. While in the "parent" i set interface, where the computer is located, in this case "to_Nick" everything works fine, the speed of torrent download is about 80Mbit/sek, no delays. As soon as i set "global" download rate drops to 10 - 15 Mbit / s and greatly increase delays, and the work is not possible, although the maximum speed of the rule has not changed.

In attachment 2 screenshots - with the "parent" option is "global" and interface ...

How soon will these updates are available? Thanks!

When will RC8 testing begin? Keen to test as I am running one of these in semi-production.

Also, the temp seems high... It's in a server room @ 16c. FYI my CCR seems to have no rear fans? Is this normal?

CCR seems to have no rear fans? Is this normal

Your temp is OK. If you can't see the fans, doesn't mean they are not there. They are inside the wind tunnel. If you open the case, you will see the tunnel. Don't they make any noise? Maybe try to load it, when CPU load increases, fans should get louder.

No worries. I will check that at work tomorrow. It's not very loaded now ~1%. CCR is much more stable on RC7

your cpu is in threshhold range when FAN spins up, so expect to see these CPU temperatures most of the time

RC7 currently Firewall doesn't have optimization for 36 cores, we will deliver that as soon as possible, and your performance should increase some 25-30 times.

Normis did you have any eta for this ?

RC7 currently Firewall doesn't have optimization for 36 cores, we will deliver that as soon as possible, and your performance should increase some 25-30 times.

Normis did you have any eta for this ?

It is hard to say, but I think in February it will be available.

Any update on when the new IPv6 routing package will be available ?

We really need route marking in IPv6.

Nick.,

RC7 currently Firewall doesn't have optimization for 36 cores, we will deliver that as soon as possible, and your performance should increase some 25-30 times.

Normis did you have any eta for this ?

It is hard to say, but I think in February it will be available.

Thanks to everyone at Mikrotik for working so hard on RouterOS 6.

It will be good to test the new routing and firewalling.

I received my CCRs the other day to replace our 2 RB1100s which cannot manage sustained data flows above 300Mbps on our 1Gbps inter-site link.

I restored the config from my 1100 to the CCR on the first one after upgrading to ROS6rc7 and it all looks good, will switch out in the morning and see how it goes.

The problem. While configuring the second CCR, I did it in a different order. Restored the 1100 backup, fixed the ports and assigned VLANs etc to teh correct interfaces, and rebooted. All good. THEN I appllied the RC7 update for ROS6 and it undid all the changes I made. Reassigned 155 IP addresses to their interfaces is not something I enjoy doing twice.

Is this an ROS6 bug? Should I report it somewhere? Have I done something silly?

Thanks in advance.
Simon

moving configuration via backup files is not recommended, please use the new "/export compact file=name" function to reliably move config. I suggest you try that. Backup file is meant to be restored on the same device only.

Thanks Normis. I will try that in future, sounds a lot less time intensive and not a feature I have ever heard about before.

I'm trying to setup the CCR1036-12G-4S using the SFP ports but no matter what settings (auto-neg on/off, fixed speeds and duplex, sfp-rate) or gbics I throw at the ports, I'm unable to get a link in the other end. The CCR seems to think it has a link but the other ends don't agree. I've tried this on rc6 and rc7. Plugging the gbic into an RB2011LS (with v5) and it links up correctly right away.

Any ideas of what could be wrong?

moving configuration via backup files is not recommended, please use the new "/export compact file=name" function to reliably move config. I suggest you try that. Backup file is meant to be restored on the same device only.

Hi Normis

Just to experiment, i reset one of my CCRs and exported my 1100s config and tried to import on the CCR. It gave me the following.

[admin@MikroTik] >> import export21Jan.rsc
"Opening script file export21Jan.rsc
Script file loaded successfullyexpected end of command (line 318 column 56)

[admin@MikroTik] >>

A colleague mentioned that he had a similar problem the other day export config to an identical 1100.

The specific bit of the export mentioned is below:

Line317 add local-address=10.10.10.250 name=openvpn-profile remote-address=\
Line318 10.10.10.249 use-compression=no use-encryption=yes use-ipv6=no use-mpls=\
Line319 no use-vj-compression=no

Any ideas?

Thanks
Simon

@simonfishley
Perhaps IPv6 package is disabled.

Simon, as Becs said, compare if the same packages are enabled in both systems. Some configuration can't be enabled if a required package is missing. Also make sure that you use the "compact" export on the first system.

The first thing he noticed when loading 100%. RC8.

The first thing he noticed when loading 100%. RC8.

Please send us a supout.rif file from your device

The first thing he noticed when loading 100%. RC8.

Please send us a supout.rif file from your device

Sent to support.

The first thing he noticed when loading 100%. RC8.

Oleg, hello! Please tell me how you use your CCR? What services are using (NAT, firewall, BGP), and how much traffic passes through it?
Олег, здравствуйте! Скажите пожалуйста, как вы используете Ваш CCR ? Какие сервисы используете(нат, файрвол, бгп) и какой объем трафика проходит через него?
Спасибо!

The first thing he noticed when loading 100%. RC8.

Oleg, hello! Please tell me how you use your CCR? What services are using (NAT, firewall, BGP), and how much traffic passes through it?
Олег, здравствуйте! Скажите пожалуйста, как вы используете Ваш CCR ? Какие сервисы используете(нат, файрвол, бгп) и какой объем трафика проходит через него?
Спасибо!

Hello. So far, the use as NAT, firewall, and queues. On average it from 200 to 300 users. Loading of 10%. Traffic is about 150 MB / s. But do not let some of the features which are stated above. On x86 i5 4 core RouteOS version 5 steeply from 400 to 500. The license will not allow it. CPU load of about 20%. Loading volume about 250-300 MB / s.

When will RC8 testing begin?

Today

v6.0RC8 DHCP server still status offered

The first thing he noticed when loading 100%. RC8.

where download RC8 ? thanks

Is there a changelog for RC8

The first thing he noticed when loading 100%. RC8.

where download RC8 ? thanks

http://www.mikrotik.com/download/share/ ... 6.0rc8.npk

This is the second day as I installed CCR-1036-12G-4S in my network.

I first had RouterOS installed on PC-s, after that I upgraded to RB1000, after that I upgraded to PowerRouter732, after that I upgraded to Axiomtek NA-822 and now the Cloud Core Router. Here is a graph of the CPU load on the Cloud Core Router:
I expected more from this Router but it looks like there is still work on the RouterOS in the way how does it handle the CPU load.

The CPU load on the previous hardware I had, Axiomtek NA-822(with dual core CPU) had peaks going to 80% CPU load and on this router they go up to 60%.

Those peaks in the CPU load from time to time are strange and they exist in every hardware that has multi cpu.

zlatkomajstor, there are some things that we are still working on. specifically:

- Speed improvement for Simple Queues (multi core optimisations, architecture optimisations)
- Speed improvement for Filter rules (most importantly some specific matchers, especially noticable on L7 rules)
- encryption (IPsec PPTP etc.) hardware acceleration support
- additional Fast path support (for bridge firewall and ip firewall)
- improved multi-threading for PPP Servers.

When these are done in the next month, you will see big speed improvements.

I hope that there will be improvements in the speed performance normis.

I use this router only for BGP, some static routing and couple of firewall rules.

I will need 2 more pieces for upgrading the rest of my core network but I will have to see results first.

I really like Mikrotik devices and software and I would hate to move to something else but I have about 6k users and I need devices that can handle that load.

hi normis


We bought a CCR 1016-12G, it have a few problems. It always restart twice a day and drops many pppoe connections during the day. we use v6.0 rc7 .do you know how can we fix these problems? thanks for help
and sorry for bad english:)

hi normis


We bought a CCR 1016-12G, it have a few problems. It always restart twice a day and drops many pppoe connections during the day. we use v6.0 rc7 .do you know how can we fix these problems? thanks for help
and sorry for bad english:)

install RC8 from the link above, and check again, if the problem persists, make a supout.rif file when the disconnections happen, and email it to support. we will check what happens, and if it's a pppoe problem or a CCR problem.

thanks

Does it means that CCR not suppurted jumbo frames?(

Does it means that CCR not suppurted jumbo frames?(

Hi!
How much traffic goes through CCR?

Does it means that CCR not suppurted jumbo frames?(

Hi!
How much traffic goes through CCR?

Summary aggregated traffic is 1 - 1.2 Gbit/s
I`m worried about interface between CCR & NAS (x86 router on Intel i7, Intel Pro/1000PT NIC) - Now 400 Mbit/s and 50 000 pps goes through this link.... and i`d like to use jumbo frames to lower the cpu utilization...

CCR 1016-12G with soft pre-6.0-rc7, 6.0-rc7 and 6.0rc8.

Problem: When using winbox open Qeueu menu, CCR reboots and log is the record: you reboot the router without proper shutdown. The same is true if the consoles I give the command: / queue> monitor.

Supout.rif file was sent to the support.

I haven't put RC8 on the on our CCR yet, but I did put it on of my test routers and noticed that the percentage sign is gone from the CPU at the top of Winbox. I'm not even sure how I noticed it.

I haven't put RC8 on the on our CCR yet, but I did put it on of my test routers and noticed that the percentage sign is gone from the CPU at the top of Winbox. I'm not even sure how I noticed it.

check for newer build, it is already resolved there. At least on linux witn wine-winbox it shows '%' sign correctly.

yep, percents are fixed =)

Hello,

I have CCR1036-12G-4S and I am testing it 3 weeks, installed RC6, RC7 and RC8 on it but one problem already exist. When I reboot, after a few hours I am trying to log in via winbox, I am in but there is no interfaces loaded. After reboot everything is back to normal, but after some hours when i log in again there is no interfaces. In IP-> Addresses it says addresses are on Unknown interface.

I have about 400 clients which are connected via PPPOE. I have 2 WAN's and I am doing load balancing with routing marks.

Other problem is related to clients with PPPOE connection. At random time there is clients dropping, random clients, random time. I noticed that there is pppoe client which status is Dynamic but NOT running and is connected. When i disconnect him, the next client which connect has the same status and hasn't internet connection 0 packets/bytes.

Last time when I tried to do supout file it failed.

Please help.

Thanks.

Hello,

I have CCR1036-12G-4S and I am testing it 3 weeks, installed RC6, RC7 and RC8 on it but one problem already exist. When I reboot, after a few hours I am trying to log in via winbox, I am in but there is no interfaces loaded. After reboot everything is back to normal, but after some hours when i log in again there is no interfaces. In IP-> Addresses it says addresses are on Unknown interface.

I have about 400 clients which are connected via PPPOE. I have 2 WAN's and I am doing load balancing with routing marks.

Other problem is related to clients with PPPOE connection. At random time there is clients dropping, random clients, random time. I noticed that there is pppoe client which status is Dynamic but NOT running and is connected. When i disconnect him, the next client which connect has the same status and hasn't internet connection 0 packets/bytes.

Last time when I tried to do supout file it failed.

Please help.

Thanks.

Are you using IP or MAC address in the Winbox ? Use IP.

I am always using IP. Tried from Local network and also remote network. I have high internet connection.

"thegreatone" do you use radius server for authenticating? because we have the same problem but i tought that the radius server has a trouble not the ccr!

i heard from my distributor the it will be a new mikrotik version just for the ccr, is it true? and it will solve all the problems!

Yes, I am using radius server. But previously I had an x86 machine with ROS 5.22 and it worked fine.

Yes, I am using radius server. But previously I had an x86 machine with ROS 5.22 and it worked fine.

+1!

At first i try to use CCR as NAS server and have the same problem... But i use PPTP connections in my network.
As soon as the number of VPN became more than 200-300 pc`s and traffic exceeded 200 Mbit/s the strange things began to happen...

In "Interfaces" is empty or shows only 10-15 tunnels of 200-300...
At random time there is clients dropping, random clients, random time too...

I use IP address in the Winbox...

Hi,

bug report in 6.0RC8 of 25/1/2012 16:30 download.

Installed 16Gb RAM Kingston KVR1333D3S9/8G * 2
the memory is detected correctly everywhere, but on the LCD 3.6Gb left only on screen...

Router seems to run correctly...

thanks

Hi,
Is it normal that In ressources IRQ, the fancon is generating about 1100 irq per second ?

6.0RC8 25/1/2012

Where I can download RC8? Very much troubles with RC7 http://forum.mikrotik.com/viewtopic.php?f=1&t=68771
Need help....

Dear Mikrotik employees,

if there is need I can give you read access to my CCR1036-12G-4S router. As I mentioned before it is working with about 6k users so you can see how does it handle that load. If you think that access to my router can help you to improve RouterOS you are more than welcome to contact me.

I have CCR1016-12G

at first when trying to past the layer7 rules from my old x86 PC to CCR it takes about 30 sec to write each line ( about 80 lines).
the other rules were pasted normally

When I reboot, after a few minutes I am trying to log in with winbox via ip , but there is no interfaces loaded. After reboot everything is back to normal, but after a while when i log in again there is no interfaces. and in ip addresses are on Unknown interface.

I have about 1500 clients which are connected via PPPOE. I have firewall, NAT, Mangle, and some of layer7 rules...

Other problem is related to clients with PPPOE connection. At random time there is clients dropping, random clients, random time. and the max no. of users that can be connect are about 250 !!!

I think there is a major problem in CCR as used for PPPoE connections !!.

Dear zlatkomajstor ..

please if you let me log in to your CCR with thanks ...

Best regards

Hello !

Here there we has this sweet router of great imbaness CCR1036-12G-4S running 6.0rc7 in this nice production environment of ours. For our needs we see it as a good fit, ceiling cat is pleased. We can has this problem now - CCR stops responding to winbox, telnet or ssh after some while, thus we ordered them gnomes to reboot the device in the nite time when sun has went to sleepa and sane people too. Of course, while gnomes reboot it quietly and almost noone notices, I find no honor in such a reboot.
Anyone else experiencing such a thing ? I reads here about this 6.0rc8, maybe any goodie fix there ? Where can I acquire this rc8 ?

Hello !

Here there we has this sweet router of great imbaness CCR1036-12G-4S running 6.0rc7 in this nice production environment of ours. For our needs we see it as a good fit, ceiling cat is pleased. We can has this problem now - CCR stops responding to winbox, telnet or ssh after some while, thus we ordered them gnomes to reboot the device in the nite time when sun has went to sleepa and sane people too. Of course, while gnomes reboot it quietly and almost noone notices, I find no honor in such a reboot.
Anyone else experiencing such a thing ? I reads here about this 6.0rc8, maybe any goodie fix there ? Where can I acquire this rc8 ?

I have lurked far and long to acquire this for you. Treasure it with all you have:
http://www.mikrotik.com/download/share/ ... 6.0rc8.npk

Does cloud core support MSDP? I don't find MSDP in cloud core v6.0rc7 :/

Does cloud core support MSDP? I don't find MSDP in cloud core v6.0rc7 :/

as far as I know RouterOS does not support MSDP. However if you configure your multicast network as IGMPv3 you can have IGMPv3 group joins for specific source. Maybe you could find this satisfactory.

Hi all,
I have a strange issue with the CCR1036-12G-4S CloudCore ....
It just randomly restarts .... 9 times today. RouterOS v6.0rc7 And the logs does not say anything ... Any ideas how to debug that issue ? Can it be related to the traffic because there was no traffic Saturday and Sunday and it was OK, but today it started again.

Does cloud core support MSDP? I don't find MSDP in cloud core v6.0rc7 :/

as far as I know RouterOS does not support MSDP. However if you configure your multicast network as IGMPv3 you can have IGMPv3 group joins for specific source. Maybe you could find this satisfactory.

IGMPv3 isn't suitable for my soluton :/ Thank you for information.

Yes, I am using radius server. But previously I had an x86 machine with ROS 5.22 and it worked fine.

+1!

At first i try to use CCR as NAS server and have the same problem... But i use PPTP connections in my network.
As soon as the number of VPN became more than 200-300 pc`s and traffic exceeded 200 Mbit/s the strange things began to happen...

In "Interfaces" is empty or shows only 10-15 tunnels of 200-300...
At random time there is clients dropping, random clients, random time too...

I use IP address in the Winbox...

I might add that not only interfaces disappear in winbox, but in the terminals (telnet and ssh). This issue was raised long ago, and I even put out screenshots. But why is it still is not solved. Just stop working SNMP.

By the way, why in version 6 profile removed the ability to view each core? So you could see on the actual kernel that is distributed.

Yes, I am using radius server. But previously I had an x86 machine with ROS 5.22 and it worked fine.

+1!

At first i try to use CCR as NAS server and have the same problem... But i use PPTP connections in my network.
As soon as the number of VPN became more than 200-300 pc`s and traffic exceeded 200 Mbit/s the strange things began to happen...

In "Interfaces" is empty or shows only 10-15 tunnels of 200-300...
At random time there is clients dropping, random clients, random time too...

I use IP address in the Winbox...

I might add that not only interfaces disappear in winbox, but in the terminals (telnet and ssh). This issue was raised long ago, and I even put out screenshots. But why is it still is not solved. Just stop working SNMP.

Also tested web interface, same problem exist in webfix.

Noticed that it happens (usually) when I have more than 200 clients connected.

I would say always whan more than 200 VPN`s....

Today I tried again to use the CCR as a NAS server:

50 vpns - works great
100 - works
150 - winbox works slowly
200 - interfaces are beginning to disappear.. (in "PPP - Interface" 5 - 15 pc`s, at the same time "PPP - Active connection" - all present)... Ethernet interfaces are not displayed...
> 250 - other clients can`t connect to NAS... log is empty... in "PPP - interface" next picture:
D - <pptp-0>
D - <pptp-1>
D - <pptp-2>
D - <pptp-3>
D - <pptp-4>
D - <pptp-5>
D - <pptp-6>
.... etc.

Max number of active vpn`s that i can see - 263.

Than i connect my x86 NAS on intel i7 (ROS 5.22) -
online 650 vpn`s, traffic 300 Mbit/s - works great....

Can't remember where I posted this previous, but bounded interfaces are broken on CCR RC7.

When making changes to bounded interfaces, they will go down and up in the logs, but they don't actually come up. A reboot of the CCR is required.

Hi all,
I have a strange issue with the CCR1036-12G-4S CloudCore ....
It just randomly restarts .... 9 times today. RouterOS v6.0rc7 And the logs does not say anything ... Any ideas how to debug that issue ? Can it be related to the traffic because there was no traffic Saturday and Sunday and it was OK, but today it started again.

Our CCR1036-12G-4S also experiences random restarts...

jan/28 02:39:14 system,error,critical router was rebooted without proper shutdown

I would like to know from MT whether this is a known bug in the software or hardware please.

jan/28 02:39:14 system,error,critical router was rebooted without proper shutdown

If it gets the date correct in the log, it is probably one of the phantom reboot log messages. The router probably did not actually reboot. Unless, do the CCRs have an actual real time clock? That would be nice. I'm not lucky enough to have hands on with a CCR yet.

My RB1100AHx2 shows reboot time at unix epoch when it actually reboots. It shows correct date/time for phantom reboot log messages. You might also look for ether down/up messages around the reboot log message. They will probably be in the log if it was a real reboot, and not if it wasn't.

there was a change that made reboot messages to be shown with correct data/time if SNTP client synced within set timeout.

It's a bugs with bandwitch limit on Queue Tree. I have started BandwitchTest between two CCR1036-12g-4s and seen next results (first number- limitation, 2nd- real bandwitch):
10 mbps - 10.1 mbps
50 mbps - 50,4 mbps
100 mbps - 101,2 mbps
150 mpps - 165 mbps (!)
various between 160 and 200 - 200 mbps (shaper did not react to changing limits- banwitch still was 200 mbps) (!)
250 mbps - 275 mbps
300 mbps - 345 mbps (!)

System- ROS6.0rc7.

HTB "global" (similar to v5.x "global-in" and "global-out") doesn't know anything about Ethernet header, so it is not calculated into limit. You can use interface HTB - they know Ethernet header and takes in into account. Each packet +18 bytes (at least)

HTB "global" (similar to v5.x "global-in" and "global-out") doesn't know anything about Ethernet header, so it is not calculated into limit. You can use interface HTB - they know Ethernet header and takes in into account. Each packet +18 bytes (at least)

It's no effect:


Apparently, the shaper is not capable of running at speeds above 100 Mbps. Here's another test, which clearly shows that the restriction of 265 Mbps actual throughput of over 330 Mbps


And the strange thing, with a limit of 330 Mbps throughput is 330 Mbps, and with a limit of 335 Mbps bandwidth of 500 Mbps for some reason

Asket, please email support@mikrotik.com about Queue Tree problem with attached supout.rif file.

Asket, please email support@mikrotik.com about Queue Tree problem with attached supout.rif file.

Done.

Hi!
I have problems with my new CCR1016 V06rc7. I've send a report to support but no answer yet.
Routes learned when i was running pppoe-server stays stucked in routing table as DAC even i stopped pppoe-server, ospf and everything else. Gateway for these routes are unknown and state reachable. The ips stucked there are not active in my network. I tried to put a routing filter and check gateways with ping but these routes are still unchainged.
My first big problem (excepting missing statistics when trafiic is more that 150 mbps and else ..) is that i have another pppoe-server with ospf and clients connected there are not working because ospf from this mikrotik announce ips as directly connected to it.

Please help!

Andrei Traian

We received our first CCR-1016 in yesterday. Got a chance today to play around with it. Overall quality of the product is higher than that of the 1100 series. The molex plug in the back of the unit is reworked, and the internal power supply is much better as well. Definitely a more solid unit. I have dealt with some Forti-Net units, this feels closer to some of those, at a MUCH lower price!

I noticed the CPU runs a higher temp than our 1100ahx2, which is understood. More under the hood. Actually the board runs a bit warmer also. So far here is only only feedback. Maybe Mikrotik has already looked into this, but normally running air horizontally across the board/heat sink usually proves to be a more efficient means of cooling. So each end of the appliance would need to have vents/holes in it, but you are moving much more air across the unit that you are than just simply having a wind tunnel on the CPU.

Just some food for thought. All in all so far so good. Our unit came with RC5, when first just logged into the unit it was discarding TX packets. Upgraded to RC8 and that issue went away.

What is the best way to move a config our 1100ahx2 to the CCR? is the /export the best way? If so, our export file isn't exporting correctly, we are having to go back and fix errors in the export command.

/export compact

will be a better way imho.

Note: Starting from v6rc1 "export compact" is default behavior. To do old style export use export verbose

http://wiki.mikrotik.com/wiki/Manual:Co ... Management

Hello.
I have one CCR1036-12G-4S as border in datacenter. Last 3 days it freezing 5-10 times in day. Watchdog reboot it. I tried rc7 and rc8 - no changes.
What I can do?

Hi, today I tried set the CCR instead of my CORE router. After CCR was started he goes on reboot every 2-3 minutes. Please notice that same that CCR was connected on network for about 20 days, without load on him and that I didn't had any problems with him.

jan/02/1970 01:01:15 system,error,critical router was rebooted without proper shutdown

v6.0rc8

ropebih, LZiX

1) Please use the latest 6.0rc8 build, which was made withing an hour, all known reported issues with CCR were fixed there,
http://www.mikrotik.com/download/share/ ... 6.0rc8.npk

2) Let us know about the results.
Please contact support output file with attached supout.rif file, if you notice any problem with CCR. It would be great to get more detailed information about "freezing".

Disappearing interfaces is already not fixed in last build!!


The build from 28 Jan worked fine for me. I am downgrading to this release.

Also, time and date is reset every time I reboot / upgrade router.

thegreatone, please report your problem to MikroTik support.
We will try to fix the problem as soon as possible.

Also, time and date is reset every time I reboot / upgrade router.

CCR does not have internal memory to save clock information, you have to configure NTP client on the router.

So we did some more testing today. Couple of things. The /export compact file=back.rsc file from our 1100ahX2 didn't completely /import onto the CCR. The firewall rules, NAT rule and graphic rules did not load for some reason, even though they were included in the rsc file. Not sure if this is a known issue or not, or normal operation, but I assume it's not normal. We are running 6.0 RC8 on the CCR.

Also, it's my understanding that all ports on the CCR are tied direct to the CPU. My question is this, are the ports tied randomly or is eth1 tied to core 0, eth2 to core 1 and etc? When looking at cpu utilization, there doesn't seem to be rhym or reason to which core gets hit and etc. Does certain traffic go to a specific core or no?

So far we have not done any stress testing.

Thoughts on the above?

sergejs, problem not fixed. I sent you supout, please check.

A few problems I have picked up:

Last night we installed 2 x CCR1036 between our DC and a fibre connected highsite. This morning one of the CCR's is frozen. Shows up in IP neighbors but cannot be accessed otherwise. Waiting for a technician to get onsite so that we can reboot and upgrade to RC8.

Aside from that there is a strange issue when running bandwidth test. Testing between the fibre connected ports between the routers I get 998mbps/998mbps up/down when testing udp, so no issue there. With TCP I don't get more than 500mbps any one direction and about 300/300 full duplex. Stranger still is that the bandwidth will then start gradually dropping down, eventually getting down to about 100mbps. Through all of this the CPU is not breaking a sweat, never gets above 4%, so that is not the problem.

I have a request to Mikrotik: could you please concentrate on stability instead of features / performance improvements. Doesn't matter how much bandwidth it can push if it cannot remain online for longer than a few days. Today both my CCR and a RB1100 running 6rc7 (official one) froze at 1:28 GMT time. The CCR came back online after 15 minutes of being down and crashed again 2 times later before I had to drive to the datacenter and put it behind a managed power strip to cut power if it does it again. The RB1100 is still bricked at this point. Tried yanking the power, but nothing - does not boot up any more. Will try to recover it via the serial cable. All the devices were doing is static routing, no queues, firewall rules, just basic static routing with 1 default route.

Please also do not call these versions RC as it is misleading at best. Anyone who used RC from Microsoft or Apple expects RC to be an actual Release Candidate as in ready to be released if no show stopper bugs are found. The RC7 or routerOS is closer to alpha 7. It would be great if you would do this in the future as the RC name was the only reason I put them in production in the first place (found this thread only later).

One side is now freezing up constantly. In addition pings across the fibre are 0ms but every 10th ping or so it either times out or jumps to 50+ms.

Hi, CCR is still rebooting very often without any good reason. I send supout files to support@mikrotik.com

Deleted because not related.

Hello savagedavid,

One side is now freezing up constantly.

This morning one of the CCR's is frozen.

And what is with the other CCR? Is it running fine or there are also problems?

We eventually figured out the freezing side had not taken the RC8 upgrade (corrupt file upload) so we fixed that. The other side running RC8 has been fine since install, aside from the odd latency issue mentioned above

Aside from the problems mentioned above when the routers are working they are properly fast! Whatever you throw at it the CPU never hits 5%. So good hardware for sure, just require a less buggy ROS release to take full advantage.

Seems I spoke too soon. The RC8 unit is still freezing. We have installed the old RB1100 at one location and continue to monitor.

Deleted because not related.

I really think that we all should change the former thinking in my opinion,
because today we have 36 cpu cores and not only one or two cores! Am I right with this?
If the CPU load from the CCR is under 36% of total CPU load it will be shown as 0-1% of each core > 36% : 36 cores = 0%.
And now backwards please, if then all CPU cores are running perhaps on 5% of usage should that mean then, we have
5% x 36 cores = 180% and are the 80% then the benefit from this device, or have I done now a thinking false?
Because in normal terms and conditions you will only and even talk about 0% - 100% right?

Okay now my head just hurts!!

I suspect we may just have a faulty unit since the other installed device is running without problem. Fortunately I have a spare which I will install tomorrow morning and monitor over the weekend.

Hello again savagedavid,

just require a less buggy ROS release to take full advantage.

this is more related to the new hardware platform type from Tilera, but if they get it false free running
and other well known bugs and issues are also fixed, it must be a really cool device, i personally think.

No I get that. Maybe we give MikroTik a hard time too quickly! We always want the very latest features as soon as possible completely bug free. I guess you can choose any 2 of those options, but not all 3!

Deleted because not related.

I think I have tracked the problem down. Seems like the CCR's are much more sensitive to heat. The CPU temp was exceeding 75 degrees in the location it was installed (the RB1100AH did not seem to have a problem in the same location). Back in the DC it does not freeze at all. I am going to look at additional cooling, but perhaps MikroTik can modify the ROS software to have more aggressive fan activation/cooling for these units.

I think I have tracked the problem down. Seems like the CCR's are much more sensitive to heat. The CPU temp was exceeding 75 degrees in the location it was installed (the RB1100AH did not seem to have a problem in the same location). Back in the DC it does not freeze at all. I am going to look at additional cooling, but perhaps MikroTik can modify the ROS software to have more aggressive fan activation/cooling for these units.

Actually its freezing even in a DC environment with 20C ambient temp. The CPU temp never exceeded 54C (average over 4 days was 53.3C) and it still freezes. Even with the yesterday's version of RC8 it only lasted 16h and then died again. Temperature may be a factor, but not the main cause.

i have a lot of problem with ccr especialy with radius server its not active well with rad control
i used another one as pppoe server its freezing very day i should reboot it also i cant see the interface when there is load more than 9% i lose all the graphics and no ppp clients are visible

Just paste that line in cli and see which item goes red - something will not be recognized nd you can ammend the script.

I think I have tracked the problem down. Seems like the CCR's are much more sensitive to heat. The CPU temp was exceeding 75 degrees in the location it was installed (the RB1100AH did not seem to have a problem in the same location). Back in the DC it does not freeze at all. I am going to look at additional cooling, but perhaps MikroTik can modify the ROS software to have more aggressive fan activation/cooling for these units.

In case of overheating your CCR will reboot itself and have 1000Mhz, if still overheating another reboot and - 800Mhz, and so on until it reaches 200Mhz. I tried it to check how board is behaving when fan dies. As soon as you reboot it manually it goes back to 1200Mhz

Dear Mikrotik employees,

if there is need I can give you read access to my CCR1036-12G-4S router. As I mentioned before it is working with about 6k users so you can see how does it handle that load. If you think that access to my router can help you to improve RouterOS you are more than welcome to contact me.

i wander how can you do this with ccr1036, could you please tell us? thanks

Hi, is it possible to connect CCR1016-12G to 220V and PoE at the same time to achieve power redundancy in case the internal power supply dies? (Similar question was asked in this thread but not answered.)

Deleted because not related.

Hi all

Can anyone comment on the stability and performance if using the ccr in a P role running LDP, OSPF and around 96+ QoS queues with DSCP.

Hi, is it possible to connect CCR1016-12G to 220V and PoE at the same time to achieve power redundancy in case the internal power supply dies? (Similar question was asked in this thread but not answered.)

@beamer
Perhaps interesting for you? http://forum.mikrotik.com/viewtopic.php?f=3&t=69422
Ok he was not telling around about witch CCR model he is talking about as I see it, but otherwise good to know, or?

I was talking about the CCR1036, but im quite sure the 1016 has also two power sockets. But he would see it as soon as he opens the case.

Dear Normis!

Sorry about the duplicate post.

Multicast routing (PIM-SM) same source (IPTV channel) with different client IP (Set Top Box) not working in CCR1036 (RC6,RC7,RC8). Same ROS version (RC6,RC7,RC8) and same configuration (1. RP, 3. Interface) working correctly in RB2011.

Please investigate this problem...


Thx.

Attila

Hello, for everyone. I want know does CCR1036-12G-4S routerboard support interface grouping? If i want 4 ports work as switch ports. Thank You for answer.

Hello, for everyone. I want know does CCR1036-12G-4S routerboard support interface grouping? If i want 4 ports work as switch ports. Thank You for answer.

Yes, simply set up a bridge and add the wanted ports.

hi, everyone^_^
i come from china
i doesn't speak english very well
could you please tell me why there are so many blanks in IRQ?
thanks a lot!

Hallo
Is there someone kindly post picture of a tcp bandwith test on more than 3 gbit ports with port mtu 1500 on ccr?
Better if used iperf or traffic generator

Reason of the question is looking for to replace a bgp router connected to internet with single NIC port and today i cannot see more than 300 Mb tcp with a x86 hw.

Thanks in advance
Massimo

Deleted because not related.

Thanks Dobby
I noted to implement on my lab test.

btw I still not seen a serious comparison between a CCR and X86

Deleted because not related.

Hello,

I didn't find an answer to this question - maybe my fault:

Which s0dimm moduls are working?
Is there a list of working components?
The list of the sfp moduls i found....

I want to use the ccr as hotspot gateway with 16 GB RAM as webcache also.
Some experience?



thanks,

Paul

Hi all,

We bought two CCR1036, both at about the same load and use the same function, but one of them has 21 days uptime (ros 6.0rc7), and another 3-4 times a day at random freezes(6.0rc7-8-9). at the time of freezing, we can not get access to the device through ssh \ telnet \ winbox, routing through the device works, but not consistently, and with a noticeable delay. the only thing we can do at this point, reboot it remotely through snmp. We mail a lot of supout files, but still have not received an answer to the question, what is the reason our problem, what we need to turn off the function to solve the problem temporarily.
We have [Ticket#2013020666001015]

with best regards

Hi all,

We bought two CCR1036, both at about the same load and use the same function, but one of them has 21 days uptime (ros 6.0rc7), and another 3-4 times a day at random freezes(6.0rc7-8-9). at the time of freezing, we can not get access to the device through ssh \ telnet \ winbox, routing through the device works, but not consistently, and with a noticeable delay. the only thing we can do at this point, reboot it remotely through snmp. We mail a lot of supout files, but still have not received an answer to the question, what is the reason our problem, what we need to turn off the function to solve the problem temporarily.
We have [Ticket#2013020666001015]

with best regards

We have your ticket, and are looking into this issue.

Deleted because not related.

you can also use higher frequency for the DDR3 module, it it backwards compatible. So DDR3 DIMM with 1800MHz will also work,

Hello. I write from Italy and I would like to have your opinion about CCR1036-12G-4S.
I wish to use it in the net of a paper mill. It will be the router for 4 switch.
On each switch there are about 10/11 PLC . Each one is an IP.
What about using it in industrial automation?
Thank you in advance

Speaking of memory.

I've got a CCR running and the memory status has gone from 3734MiB to 2990.7 MiB in a matter of 3 days. Is this expected? Normal?

Speaking of memory.

I've got a CCR running and the memory status has gone from 3734MiB to 2990.7 MiB in a matter of 3 days. Is this expected? Normal?

It depends on what configuration you have. Some features do use memory.

Deleted because not related.

you can also use higher frequency for the DDR3 module, it it backwards compatible. So DDR3 DIMM with 1800MHz will also work,

Haha, super thanks for this hint normis!
But do they accept ECC/REG modules would it better to go with the normal S0DIMM?
For sure it is the first series and I would expect to much at first, but good to know which modules are
working "not" is also very welcome.

We have not specifically made ECC support, so I can't say. It might work, but it will certainly not use the ECC features

Deleted because not related.

Hi all,

We bought two CCR1036, both at about the same load and use the same function, but one of them has 21 days uptime (ros 6.0rc7), and another 3-4 times a day at random freezes(6.0rc7-8-9). at the time of freezing, we can not get access to the device through ssh \ telnet \ winbox, routing through the device works, but not consistently, and with a noticeable delay. the only thing we can do at this point, reboot it remotely through snmp. We mail a lot of supout files, but still have not received an answer to the question, what is the reason our problem, what we need to turn off the function to solve the problem temporarily.
We have [Ticket#2013020666001015]

with best regards

hi simpl3x,

we had same problem here. routerboard was shipped with 6.0rc6 and we make upgrade to 6.0rc7-9.
this was solution for us (tnx to wirac.ba mikrotik team):

1. make backup to disk
2. hard reset your ccr (hold metal object on jumper hole while the board boots)
3. restore backup from disk (if your think that your configuration is ok)
4. check if cpu load is growing and if cpu fan go on after 45-50°C

best regards
nedim

support and also my reseller didn't respond me till now...

is this module ok:

Kingston Value Ram 8 GB DDR3, 1333 Mhz (PC3-10600), S0Dimm 204 Pin , CL9

???

or has some one other advices?

Paul

So even when sodimm support ecc ros or cpu bus does not support?
Just for my understanding

Deleted because not related.

So it seems RC7 is still the most stable? On this, 14day uptime

Hello !

After upgrade to RC9, problem with Winbox still remains - after some time CCR stops responding to Winbox/telnet/SSH. There is extra info on this behavior: while all management access to router is gone, there are some interfaces and MAC adresses which DO respond to Winbox. Also, after few days of uptime, somehow router becomes accessible *again* with winbox/telnet/ssh thru any IP address it has.

Hello !

After upgrade to RC9, problem with Winbox still remains - after some time CCR stops responding to Winbox/telnet/SSH. There is extra info on this behavior: while all management access to router is gone, there are some interfaces and MAC adresses which DO respond to Winbox. Also, after few days of uptime, somehow router becomes accessible *again* with winbox/telnet/ssh thru any IP address it has.

Same problem. Router accessible only via mac-address (no ip). Problem starting with multiple ip addresses on one interface.

I have ordered 2 pcs of CCR1036 still waiting for delivery until December. But I`m not hurry because OS version is still RC. My christmas wish was to be announced stable ROS version before CCR arrives. Want to see stable CCR with ROSv6 then maybe additional features like Metarouter, encryption offload etc...

After upgrade to RC9, problem with Winbox still remains - after some time CCR stops responding to Winbox/telnet/SSH. There is extra info on this behavior: while all management access to router is gone, there are some interfaces and MAC adresses which DO respond to Winbox. Also, after few days of uptime, somehow router becomes accessible *again* with winbox/telnet/ssh thru any IP address it has.

Same problem. Router accessible only via mac-address (no ip). Problem starting with multiple ip addresses on one interface.

Stop panic, guys... this issue IMHO not only RC9, ROSv6 and even CCR.
Just now I suffer the same pain for a month with cheap RB750, and periodically with RB751U-2HnD also. The last one I was able to reboot via PoE, but the first one I can't reboot remotely. Because it handles traffic so well I leave it up working till another important reasons for reboot or need to change configuration.

Ansy, Bakula, Jetrider, please write to support with information how we could connect to your router with Winbox and test this.

Hi normis,

When is the RouterBOOT memory tester going to be implemented on the CCR-series ?

If we are going to upgrade the memory, we would like to be reasonably certain that the memory is 100% functional in the CCR before deploying it in a remote datacentre.

Regards,
Terry Froy
Spilsby Internet Solutions
http://www.spilsby.net/

When is the RouterBOOT memory tester going to be implemented on the CCR-series ?
If we are going to upgrade the memory, we would like to be reasonably certain that the memory is 100% functional in the CCR before deploying it in a remote datacentre.

Before upgrade memory in CCR --> you can make memory test in any X86 machine.

Ansy, Bakula, Jetrider, please write to support with information how we could connect to your router with Winbox and test this.

Ticket#2013021366000842
video + supout.rif + access to demo router

Ansy, Bakula, Jetrider, please write to support with information how we could connect to your router with Winbox and test this.

Ticket#2013021366000842
video + supout.rif + access to demo router

replied. looks like issue is fixed in rc10

Ansy, Bakula, Jetrider, please write to support with information how we could connect to your router with Winbox and test this.

Ticket#2013021366000842
video + supout.rif + access to demo router

replied. looks like issue is fixed in rc10

this problem Ticket#2013020666001015 fixed in rc10?

What would be the best queueing method on the CCR?
I read on this forum that on Multi-Core/CPU systems, multi queue PFIFO is the best option (performance wise).
Is this also the case on the CCR?

replied. looks like issue is fixed in rc10

this problem Ticket#2013020666001015 fixed in rc10?[/quote]

we did find some significant issue and fixed it, please try todays build:
http://www.mikrotik.com/download/share/ ... .0rc10.npk

we did find some significant issue and fixed it, please try todays build:
http://www.mikrotik.com/download/share/ ... .0rc10.npk

thank you Normis, we already installed it 10 minutes ago(Feb/15/2013 10:47:45), your debug packet was deleted before it. we mail supout files if problem continues

Hello guys , i need to enable on syncookie protection for detect and block spoof ips. I already enabled TCP-Syn cookies in ip settings and it wont help to prevent ddos attack.
Routerboard inaccessible when synflood attack comes from spoof ips. If i limit whole new synconnections routerboard drops all new and old requests.

Please help me to solve it .

nothing has changed in rc10. 5-4 routing freezes and not available ssh\winbox services a day. maximum 11 hour uptime.

nothing has changed in rc10. 5-4 routing freezes and not available ssh\winbox services a day. maximum 11 hour uptime.

could you please write to a support@miktotik.com or if you already have done that write the ticket number so we could check?

MikroTik,

Have the firewall and queue multi threading optimizations made it into the latest releases? There haven't been any specific mentions of it.

RC7 currently Firewall doesn't have optimization for 36 cores, we will deliver that as soon as possible, and your performance should increase some 25-30 times.

Our test-CCR just died. No traffic, OSPF running.
Have to powercycle. RC9. Runtime 2 days.

Our test-CCR just died. No traffic, OSPF running.
Have to powercycle. RC9. Runtime 2 days.

Please contact support@mikrotik.com to get the newest RouterOS version which will help to solve this problem.

hello,
please, check mail about [Ticket#2013020666001015], we have made access through rs232 port another mikrotik, and get supout.rif at the moment of freeze.

hello,
please, check mail about [Ticket#2013020666001015], we have made access through rs232 port another mikrotik, and get supout.rif at the moment of freeze.

Thank you, we already working on your ticket.

I don't know if it has been noted, but it doesn't seem that the UPS package is functional on the ccr with 6.0rc9.

When will METAROUTER be implement on the new cloud core router ?

We had a complete freeze. No access with serial port. Have to RMA it.

We had a complete freeze. No access with serial port. Have to RMA it.

please update your support ticket with this information so we could check what is going on with that router.

MikroTik,

Have the firewall and queue multi threading optimizations made it into the latest releases? There haven't been any specific mentions of it.

normis wrote:
RC7 currently Firewall doesn't have optimization for 36 cores, we will deliver that as soon as possible, and your performance should increase some 25-30 times.

I don't know if it has been noted, but it doesn't seem that the UPS package is functional on the ccr with 6.0rc9.

Any timeframe on when this would be remedied?

popcorrin, which port do you use for connecting UPS, serial0 or usb1?

usb1 should work without changes in existing configuration. (works for me on CCR with v6.0rc9)
serial0 by default is used by Console, it needs to be disabled in System -> Console menu. Also it might need Baud Rate change in System -> Ports.

Hii All,

I tried to make MRTG for queue tree using OID but the graphic sometime see sometime not, why?? But if i make MRTG for the interface (not OID), the graphic show up. Can anyone help...??
My CCR
uptime: 3d22h15m27s
version: 6.0rc9
build-time: Feb/08/2013 08:15:34
free-memory: 1666.2MiB
total-memory: 1940.2MiB
cpu: tilegx
cpu-count: 16
cpu-frequency: 1200MHz
cpu-load: 2%
free-hdd-space: 434.7MiB
total-hdd-space: 512.0MiB
architecture-name: tile
board-name: CCR1016-12G
platform: MikroTik
Thanks.

bug report:

RC11
LCD info slideshow, fan & fan2 to 0 rpm....

bug report:

RC11
LCD info slideshow, fan & fan2 to 0 rpm....

wait longer. it works

popcorrin, which port do you use for connecting UPS, serial0 or usb1?

usb1 should work without changes in existing configuration. (works for me on CCR with v6.0rc9)
serial0 by default is used by Console, it needs to be disabled in System -> Console menu. Also it might need Baud Rate change in System -> Ports.

Okay, maybe it's the device and not the ccr. I just assumed the ups package for the ccr needed some work.

The device is the ups-mt from http://tandem.ck.ua/ups_mtm-eng.php.
It is a serial device and I have disabled serial0 under the system>console menu.

The ccr doesn't even recognize it.

I have a ups-mt on a RB450G and it is recognized by the RB450G but it doesn't seem to be very accurate. Shows online when it is actually off line and incorrect voltages. I'm starting to get the impression that these devices aren't quite up to snuff.
I'll try to find time to test with a ups from some other manufacturer and report back.

First of all hi to everyone (this, for me being new to this forum)

I think we are more than a few who are waiting for this:

RC7 currently Firewall doesn't have optimization for 36 cores, we will deliver that as soon as possible, and your performance should increase some 25-30 times.

Any news are welcome!

IPIP tunnel mikrotik<->linux after 20-30 minutes stops passing packets.
Pings with size 50bytes pass all the time.
Packets bigger than 1300bytes stops passing, after about 5 minutes later, they suddenly start to pass again.

Changed from xeon with 5.22, where it was without any problem.

Problem si on CCR, v6, rc11, rc10 too

I estabilished eoip tunnel with emulation of eoip under linux by meoip to oveerride this problem and it works.

Okay, maybe it's the device and not the ccr. I just assumed the ups package for the ccr needed some work.

The device is the ups-mt from http://tandem.ck.ua/ups_mtm-eng.php.
It is a serial device and I have disabled serial0 under the system>console menu.

The ccr doesn't even recognize it.

I have a ups-mt on a RB450G and it is recognized by the RB450G but it doesn't seem to be very accurate. Shows online when it is actually off line and incorrect voltages. I'm starting to get the impression that these devices aren't quite up to snuff.
I'll try to find time to test with a ups from some other manufacturer and report back.

I tried the exact same device on an mipsbe platform and it seemed to work as expected which leads me to believe that it has something to do with the ccr. Anybody at mikrotik have a serial ups that they can test with?

MikroTik,

Have the firewall and queue multi threading optimizations made it into the latest releases? There haven't been any specific mentions of it.

normis wrote:
RC7 currently Firewall doesn't have optimization for 36 cores, we will deliver that as soon as possible, and your performance should increase some 25-30 times.

When finally CCR will work with all kernels. Honestly, look at the loading 4 cores on x86 nicer than CCR. It really shows the overall load of the order of 30-50% of the CCR 5-6%. When loading CCR 5-6% it starts to lags. By the way the estimated 88% of one core - is just 2.4% of the total charge of the nucleus. There is clearly seen that one process is allocated to only one core.

Oleg, what features do you use there?

Normis, same situation, as with OlegM.

I tried to use ccr1036 as pptp NAS with queues and nat.
Seems, when traffic come to around 100Mbps, some cores bring 100% use by IRQs, that may lead to lags and traffic forwarding stop.
Else processes in profile look like low CPU consumption.

[pater@ccrvpn] > tool profile cpu=total
NAME CPU USAGE
ppp all 2.4%
pptp all 0%
ppp all 0%
console all 0%
ssh all 0%
traffic-flow all 0%
firewall all 1.5%
mpls all 0%
management all 1.7%
encrypting all 0%
idle all 92.5%
profiling all 0%
queuing all 1.1%
routing all 0.1%
unclassified all 0%

Here is sample CPU load (look to cpu5 - same randomly occured with any single core with load up to 100%)

[pater@ccrvpn] /system resource cpu> print
# CPU LOAD IRQ DISK
0 cpu0 16% 2% 0%
1 cpu1 0% 0% 0%
2 cpu2 0% 0% 0%
3 cpu3 13% 4% 0%
4 cpu4 27% 13% 0%
5 cpu5 88% 88% 0%
6 cpu6 2% 2% 0%
7 cpu7 21% 6% 0%
8 cpu8 1% 1% 0%
9 cpu9 2% 2% 0%
10 cpu10 1% 1% 0%
11 cpu11 0% 0% 0%
12 cpu12 6% 6% 0%
13 cpu13 2% 0% 0%
14 cpu14 2% 2% 0%
15 cpu15 0% 0% 0%
16 cpu16 1% 1% 0%
17 cpu17 10% 0% 0%
18 cpu18 0% 0% 0%
19 cpu19 6% 6% 0%
20 cpu20 1% 1% 0%
21 cpu21 0% 0% 0%
22 cpu22 18% 4% 0%
23 cpu23 1% 1% 0%
24 cpu24 0% 0% 0%
25 cpu25 10% 0% 0%
26 cpu26 0% 0% 0%
27 cpu27 3% 3% 0%
28 cpu28 19% 0% 0%
29 cpu29 4% 0% 0%
30 cpu30 17% 0% 0%
31 cpu31 0% 0% 0%
32 cpu32 0% 0% 0%
33 cpu33 1% 1% 0%
34 cpu34 0% 0% 0%
35 cpu35 5% 5% 0%


Normis, seems bottle neck is not queue but IRQ balancing|sharing for core ethernet.
While reading forum, i often look complains about poor performance not only in queues and pptp tunnels, but also in general applications such VLANs.
Also, according profiler, seems queue process utilize much less CPU, than firewall procees, used for nat and packet marking.

RC12 will have significant speed speed increase for Queues on CCR

Problem with IP access thru winbox/telnet/ssh still remains in RC12. Just sent in another supout. Please, oh Lords-ah at MT, deliver us from dis evil plague !

Normis, seems bottle neck is not queue but IRQ balancing|sharing for core ethernet.
While reading forum, i often look complains about poor performance not only in queues and pptp tunnels, but also in general applications such VLANs.
Also, according profiler, seems queue process utilize much less CPU, than firewall procees, used for nat and packet marking.

Bring up more than 100Mbps without lags and traffic stopping is impossible for ccr now, so in the next month, we may have to purchase another vendor's equipment for prodution environment.

Oleg, what features do you use there?

VRRP, PPTP Server, Firewall (Layer 7 is not used), Queues,Traffic Flow, SNMP, Radius (login, ppp) . In my forgotten nothing.

We are looking for a replacement for our microsoft TMG server as it has been nothing but trouble as our network has expanded past 2000 users and its running on crappy windows software.

As a home user of a RB450G i think that the CCR would be perfect for the school i work for

My question is if we bought 2 CCR1036-12G-4S can we have hardware failover for just one WAN link with VRPP or similar

I know the hardware platform is more than capable as a good firewall for us, certainly much better than TMG but we would like to be safe in the knowledge that if one CCR packs up then the other would take over.

The WAN connection is from a cisco router with only one copper port output so we cant request multiple gateways from ISP. so i was thinking if the router supported bypass that in the event that it died the connection would loop through to the other router.

Anyones thoughts would be appreciated

Im having issues with TX drops on the Ethernet ports on the CCR1036-12G-3S. Changed cables still the same and there is no tx drops when connected to the old RB1100

Im having issues with TX drops on the Ethernet ports on the CCR1036-12G-3S. Changed cables still the same and there is no tx drops when connected to the old RB1100

They know about this.. I`m testing rc12 with changed ethernet driver buffer size, but is the same after 2-5 days uptime .. TX drops is showing up...

The only solution is reboot at 5-6am every 2 days

WIth Regards,
Majkel

what interface queue type do you use? what if you change it to mq-pfifo?

I try changing snd testing everything I think. pfifo limit too Now I use multi queue

v6.0rc11 with CCR1036 reboots by watchdog every few hours. When disable watchdog, CCR1036 just crushes. Occurs same issue on both 2 CCR1036s. Role is PPPoE server, some firewall rules, 4 static simple queues, radius AAA, around 200 NAT rules, which translate customers to 200 public ip dynamically. Does any body see same issue?

I don`t use watchdog now.. but I`v got the same problem with him on rc9.. And I need to configure my 1036 from the beginning..

The further away, the less work CCR from rc6 and ending rc12. Throughput dropped. When connecting more than 150 weird things happen. PPTP connection via Radius hang for a few minutes and eventually No results. If earlier when connecting to 300 users from going interfaces, now began to disappear in the Winbox rules Simple Queues. When at last you will see the normal version, as in version 5? Promised in February to eliminate the problem. And it turns out that rc worse with each release.

Hi,

cisco 2901 replaced with CCR1016-12, updated to v6.0rc11, 3 nat rules, 2 firewall rules, only 3 ethernet ports connected,

around 50Mbits traffic, only TX Drops on ethenet ports. WOW!!!!

No packets dropped with old cisco.

Hi,

cisco 2901 replaced with CCR1016-12, updated to v6.0rc11, 3 nat rules, 2 firewall rules, only 3 ethernet ports connected,

around 50Mbits traffic, only TX Drops on ethenet ports. WOW!!!!

No packets dropped with old cisco.

RC12 has improvements for this issue. Try this version: http://www.mikrotik.com/download/share/ ... .0rc12.npk

Normis,

RC12 causes router reboot when traffic reaches around 200MB. Log message says: "router was rebooted without proper shutdown".

I have one CCR1016 with 10 RB 751 plugged. When I btest from one RB 751 and traffic reaches 200MB the router reboot.

Normis,

RC12 causes router reboot when traffic reaches around 200MB. Log message says: "router was rebooted without proper shutdown".

I have one CCR1016 with 10 RB 751 plugged. When I btest from one RB 751 and traffic reaches 200MB the router reboot.

when did you download rc12? it is being updated every 30 minutes or so. try newer.
latest build was 13 minutes ago (6.0rc12 (2013-Mar-06 15:17)

90 seconds after bstest starts the router reboot.

I was testing with UDP traffic on both directions started from RB 751.

Normis,

RC12 causes router reboot when traffic reaches around 200MB. Log message says: "router was rebooted without proper shutdown".

I have one CCR1016 with 10 RB 751 plugged. When I btest from one RB 751 and traffic reaches 200MB the router reboot.

when did you download rc12? it is being updated every 30 minutes or so. try newer.
latest build was 13 minutes ago (6.0rc12 (2013-Mar-06 15:17)

After new update I saw no changes. The problem remains...

Normis,

Actually the problem is something about "management" process. I started test from 2 RBs 751 using direction=transmit and I saw no problem.
But when I started test from one RB 751 using direction=receiving "management" process reaches 6% on CCR1016 and after 90 seconds the router reboot.

Actually the problem is something about "management" process. I started test from 2 RBs 751 using direction=transmit and I saw no problem.
But when I started test from one RB 751 using direction=receiving "management" process reaches 6% on CCR1016 and after 90 seconds the router reboot.

I guess the same issue rebooted my RB750UP in switch mode when passing much transit traffic in only one direction (but OK with another direction).

[Ticket#2013022766000156] ROS v.5.24 and older.

After I reconfigured RB750UP's switch scheme (master-slave) to bare bridge on that two ports, all works OK. Please, try it if applicable.

Normis,

Actually the problem is something about "management" process. I started test from 2 RBs 751 using direction=transmit and I saw no problem.
But when I started test from one RB 751 using direction=receiving "management" process reaches 6% on CCR1016 and after 90 seconds the router reboot.

Look like Btest causes memory leak on the CCR. Thanks! We will fix it.