There’s your answer. Public IP addresses are always getting scanned. Whenever a scan requests a public IP that doesn’t exist, ROS must send an ARP request for this address - it stores the “incomplete” ARP in the ARP table. These are displayed as all-zeroes in the GUI, or just as “incomplete” in the terminal.
It’s normal behavior. You’re not seeing it on private-only interfaces because none of your internal hosts is doing IP scans or trying to reach an IP that does not exist.
Sadly, still no luck on getting 6.35 final work.
Right now it is i have my config or i have 6.35 final, if both - i have all interfaces down completely after boot finishes.
BUG !!
this things made me crazy
i have my router-board 1100 with a normal router mode for 50 IPs queue , in an aggerate of 50MB for all the internet
after upgrade , all clients including me ., (adminstrator) which i am not inside of any queue , have a very horrible speed packet loss and slow response ,
all of the customers had simultaneously very bad internet service response ,
anyhow , i had to do a Netinstall to recover for a previuos version so the service went back normal
now to the other end , HOW is it possible to alter this new queue setting "bucket size " for a dhcp queue or for a hotspot queue . !
I would venture to guess that the ros6.35 has problems with ethernet on CloudCore routers.
I have Ether1 Flapping, which only started to flap after upgrade, the router it is connected to receives no Link Partner Advertisments.
Also SFPplus1 was set to No Autonegotiate at 1gps, after upgrade it must be set to autonegotiate to work.
[EDIT]
SFPplus1 registers a connection when autonegotiate is set, but does not pass traffic.
[/EDIT]
I have had some issues with NAT on this version, had a simple NAT rule for SMTP to internal mail server and router was dropping packets on external interface. Funny thing is some connections actually went through, did investigate this because people complained about mail not being received. Have reverted to previous version for now.
You probably have a CCR and had the interfaces MTU set to a value bigger than 10222. Set it to 10222 and your interfaces will be happy and kicking again after the upgrade.
It made me waste almost one hour of time finding why the router was inoperational after the upgrade.
In /system logging I have firewall topic set to send events to a remote host and via email.
I receive the events on the remote host, but none are being sent via email.
Emailing works ok, as I receive other type of logger topics ok. Critical events are sent properly.
4 critical sendmail
7 firewall sendmail
Both critical and firewall topics are using the same action, but only critical are sent properly.
Can you please take a look?
Thanks,
Mihai.
LE : I confirm that going back to 6.34.4 fixes it.
Im tring the new FW on RB411U with MC7710 Sierra Wiereless modem and 4G SIM (but the serice 4G is not yet enable by ISP) and I notice settng 3G + LTE the interface LTE doesnt connect.
Resuming:
Using GSM + 3G + LTE the interface connect on 3G (is OK).
Using 3G only the interface connect (is OK).
Using GSM + 3G the interface connect (is OK).
Using 3G + LTE the interface doesnt connect (is not OK). It should connect to 3G (why not?).
Thanks, DaxStox for report, we will try to replicate that issue.
There is an issue with CAPsMAN in 6.35.
It could lockup wireless interface on CAP, if CAP looses connection to CAPsMAN on second retry, We are working on fix for that right now. For now, simple reboot will help till the next connection loss.
This is the first release where MT hasn’t posted a torrent file. They stopped linking to them on the downloads page, but they were still there.
So if you need torrent, download from page and create your own..
Obviously, you no longer “need” the torrent once you have downloaded ~15 links, unpacked everything and arranged it all in a directory structure… but it sure would have been a time-saver.
You are welcome. Can make request?: Make changelog downloadable, it’s primary function is documentation replacement …
[Edit] Or at least make release notes part of documentation wiki.
Its still a bug, the import script should have detected 12226 and since now only max of 12222 is supported, it should have dropped them back, thats why they disable on services starting as it disables the Ethernet ports
That’s not going to work the way it used to anymore. Now you have to use CoA to change dynamic items.
PS. Please learn how to use forum search function. This question has been asked and answered enormous amount of times already. If you need some examples- again, search the forum!
Good day.
Hardware - RB951G-2HnD
After update to 6.35 begin trouble with SNMP.
My Zabbix server can’not read parameter “Current CPU load”
OID 1.3.6.1.4.1.2021.11.52.0
Please confirm this bug and fix it in next update