Just got the update from my network manager now:

I've just had a chat with Kevin @ Netgate, apparently this morning IHUG
decided it would be an excellent idea to advertise all customers /32
subnets from eigrp into BGP (that's one hell of a lot of routes).

This of course caused an avalance of routes to spew out to Netgate and
from there on down to us (and other Netgate customers).

This explains the increased CPU load on all three routers (as they
struggled to cope with all the routes) and then the final cries as they
ran out of memory:

May 3 10:31:52 kelly 15285: 24w5d: %SYS-2-MALLOCFAIL: Memory allocation of
82064 bytes failed from 0x602401DC, pool Processor, alignment 0
May 3 10:31:52 kelly 15286: -Process= "BGP Router", ipl= 0, pid= 85
May 3 10:31:52 kelly 15287: -Traceback= 6023E4BC 6023FE38 602401E4 6023BD58
6023C424 6039C1E0 6039C400 6039D2EC 604CDDDC 604CE2C8 60915158 604CE374
604D2E78 60239B4C 60239B38
May 3 10:31:58 cassandra 11: 2w5d: %SYS-2-MALLOCFAIL: Memory allocation of
57664 bytes failed from 0x602401DC, pool Processor, alignment 0
May 3 10:31:58 cassandra 12: -Process= "BGP Router", ipl= 0, pid= 84
May 3 10:31:58 cassandra 13: -Traceback= 6023E4BC 6023FE38 602401E4
6023BD58 6023C424 608FBB1C 608FBC70 60196004 603BFF20 6039C680 6039E880
604CDDDC 604CE2C8 60915158 604CE374 604D2E78
May 3 10:31:58 cassandra 14: 2w5d: %FIB-3-NOMEM: Malloc Failure, disabling
DCEF
May 3 10:32:56 kelly 15288: 24w5d: %SYS-2-MALLOCFAIL: Memory allocation of
72000 bytes failed from 0x6023BDB4, pool Processor, alignment 16
May 3 10:32:56 kelly 15289: -Process= "BGP Router", ipl= 0, pid= 85
May 3 10:32:56 kelly 15290: -Traceback= 6023E4BC 6023FB08 6023BDBC 6023C424
601ADDA8 60187EC8 6018A104 6018DDC4 60195FDC 603BFF20 6039C680 6039E880
604CDDDC 604CE2C8 60915158 604CE374
May 3 10:32:56 kelly 15291: 24w5d: %FIB-3-NOMEM: Malloc Failure, disabling
DCEF

Bertha didn't fall over as she's got twice the RAM that kelly and
cassandra have - she coped.

According to Kevin, Netgate are moving all their connections to BGP or
static only, all BGP is also moving to extended access-list
filtering. This will prevent similar things happening in future.

> -----Original Message-----
> From: owner-adsl@ns.freebsddiary.cx
> [mailto:owner-adsl@ns.freebsddiary.cx]On Behalf Of Dan Langille
> Sent: Wednesday, 3 May 2000 11:02
> To: adsl@freebsddiary.cx
> Subject: short ADSL outage
>
>
> Today my ADSL connection died at about 10:30. At about 10:50, I rang
> Paradise. They were in the same spot. I'm back online now at 11:00.
> --
> Dan Langille [I'm looking for more work]
> http://www.langille.org/ | http://www.unixathome.org/
> http://www.racingsystem.com/ | http://www.freebsddiary.org/
>
> To unsubscribe: send mail to majordomo@freebsddiary.cx
> with "unsubscribe adsl" in the body of the message
>
>

To unsubscribe: send mail to majordomo@freebsddiary.cx
with "unsubscribe adsl" in the body of the message
Received on Wed May 3 12:18:37 2000