bird 1.5.0-4 ospf Inactivity timer expired.

[Andrey Kitsul]

Why ospf restart? Dead interval 4sec not been reached.
> 2015-10-26 17:35:11 <TRACE> kernel1: Scanning routing table
> 2015-10-26 17:35:11 <TRACE> kernel1: Pruning table master
> 2015-10-26 17:35:11 <TRACE> kernel1: Pruning inherited routes
> 2015-10-26 17:35:11 <WARN> Event 0x0000000000442440 0x0000000000000000 
> took 1873 ms
> 2015-10-26 17:35:13 <TRACE> ospf_area0: Inactivity timer expired for 
> nbr 195.x.y.z2 on eth0.1889
> 2015-10-26 17:35:13 <TRACE> ospf_area0: Neighbor 195.x.y.z2 on 
> eth0.1889 changed state from Full to Down
> 2015-10-26 17:35:13 <TRACE> ospf_area0: Neighbor 195.x.y.z2 on 
> eth0.1889 removed
> 2015-10-26 17:35:13 <TRACE> ospf_area0: New neighbor 195.x.y.z2 on 
> eth0.1889, IP address 195.x.y.z2
> 2015-10-26 17:35:13 <TRACE> ospf_area0: Neighbor 195.x.y.z2 on 
> eth0.1889 changed state from Down to Init
> 2015-10-26 17:35:13 <TRACE> ospf_area0: Neighbor 195.x.y.z2 on 
> eth0.1889 changed state from Init to 2-Way
> 2015-10-26 17:35:13 <TRACE> ospf_area0: Neighbor 195.x.y.z2 on 
> eth0.1889 changed state from 2-Way to ExStart
> 2015-10-26 17:35:13 <TRACE> ospf_area0: Neighbor 195.x.y.z2 on 
> eth0.1889 changed state from ExStart to Exchange
> 2015-10-26 17:35:13 <TRACE> ospf_area0: Scheduling routing table 
> calculation
> 2015-10-26 17:35:13 <TRACE> ospf_area0: Neighbor 195.x.y.z2 on 
> eth0.1889 changed state from Exchange to Full


26.10.2015 17:09, Ondrej Zajicek пишет:
> On Mon, Oct 26, 2015 at 04:57:08PM +0600, Andrey Kitsul wrote:
>> Hello,
>>
>> I see in the logs:
>>> 2015-10-26 16:52:11 <WARN> Event 0x0000000000442440 0x0000000000000000
>>> took 1418 ms
>>> 2015-10-26 16:52:26 <WARN> Event 0x0000000000442440 0x0000000000000000
>>> took 1523 ms
>>> 2015-10-26 16:52:41 <WARN> Event 0x0000000000442440 0x0000000000000000
>>> took 1452 ms
>>> 2015-10-26 16:52:56 <WARN> Event 0x0000000000442440 0x0000000000000000
>>> took 1429 ms
>>> 2015-10-26 16:53:11 <WARN> Event 0x0000000000442440 0x0000000000000000
>>> took 1405 ms
>>> 2015-10-26 16:53:26 <WARN> Event 0x0000000000442440 0x0000000000000000
>>> took 1428 ms
>>> 2015-10-26 16:53:41 <WARN> Event 0x0000000000442440 0x0000000000000000
>>> took 1493 ms
>>> 2015-10-26 16:53:56 <WARN> Event 0x0000000000442440 0x0000000000000000
>>> took 1455 ms
>> How to interpret?
> You would have to use gdb or perhpaps addr2line to translate addresses
> to event function names but with 15 s period it is pretty obvious that
> it is kernel periodic scan. You can confirm that by enabling debug
> { events } in the kernel proto and see kernel periodic events.
>

-- 
Кицул Андрей
администратор Unix-систем
ЗАО "НГС"
+7-923-111-3940