RB1100 random chrashing

Hardware MikroTik hardware general
1

Hi!

I am having problems with a pair of RB1100 that we are using as our main BGP routers. The problem is that they are acting very strange, they work for about a week and then just stop working, like they would shut down. When ping-eg they return Host unreachable. However the lights are on, they seem deed. After a reboot they begin to function normally for some time and after a week they crash again. Each router does that on its own (one is OK, the other crashes). Both of the routers are running metaRouter devices on them, but I believe that should not be the cause of trouble. CPU loads are low, the temp. on the CPU is about 25-30 degree C (air conditioned rack hosting place). I have also updated the routers from 4.15 to 4.17 but that did not solve the problem either. I have also tried to access the router (when it crashed) via serial console and got out this:

MikroTik 4.17
MikroTik Login: admin
Unable to handle kernel paging request for data at address 0x0000034c
Faulting instruction address: 0x8001ed74
Oops: Kernel access of bad area, sig: 11 [#6]
RB1100
NIP: 8001ed74 LR: 8001ed6c CTR: 00000000
REGS: 9fbd3df0 TRAP: 0300   Tainted: P      D    (2.6.27.39)
MSR: 00029002 <EE,ME>  CR: 88000022  XER: 00000000
DEAR: 0000034c, ESR: 00800000
TASK = 9f8aa8a0[892] 'watty' THREAD: 9fbd2000
GPR00: 8001ed6c 9fbd3ea0 9f8aa8a0 9d84e780 00000000 00000000 9d84e82c
00000000 
GPR08: 9d84e7ac 0000001f 00000000 000002ff 05000347 10024108 00715be0
00100000 
GPR16: 00000000 ffff4679 00000001 000000ff 00000000 00000000 9fbd3f50
00000000 
GPR24: 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000 
NIP [8001ed74] copy_process+0x3d4/0xbbc
LR [8001ed6c] copy_process+0x3cc/0xbbc
Call Trace:
[9fbd3ea0] [8001ed6c] copy_process+0x3cc/0xbbc (unreliable)
[9fbd3ef0] [8001f658] do_fork+0xfc/0x248
[9fbd3f30] [80005d84] sys_fork+0x54/0x68
[9fbd3f40] [8000c6ec] ret_from_syscall+0x0/0x3c
Instruction dump:
419e0038 73600400 41820018 7c001828 30000001 7c00192d 40a2fff4 4800001c 
38810008 480625b5 2c030000 4182000c <907f034c> 93810008 83810008 2f9c0000 
---[ end trace 83060596d9cc4f29 ]---

I would be grateful if someone could interpret this dump for me, so that I would know if the devices are faulty and to contact my sales provider to get a new pair of RB1100.

Thx in advance.

2

You might have more luck emailing support@mikrotik.com for an official opinion.

At a hunch I’d disable the meta router instances and see if it stabilizes. For a while that feature was so bad I could get an RB1000 to crash just by adding an instance.

3

I had issues with MetaROUTER on my RB1100. Disabled it entirely and things work well. I would definitely recommend trying that.

4

Hi bss,

Googling I found your post and I’m having the same problem here.
Did you solve the issue?

Thanks!

5

No, I tried dissabling the MetaRouter but it did not help. I have returned the devices to the supplier. I suspect that it might be some sort of RAM issue, maybe faulty memory module.