[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

bgpd: "send_rtmsg: action foo, prefix bar/baz: (some error)"



I'm trying the nice, new bgpd on for size, because I'm having trouble
with Zebra's routing tables falling out of sync with the kernel's, and
I'm getting some error messages that could indicate that the same
thing still happens.  e.g.:

Aug 10 14:04:41 btn-gw bgpd[5238]: send_rtmsg: action 1, prefix 132.203.0.0/16: 
No such process
Aug 10 14:04:41 btn-gw bgpd[5238]: send_rtmsg: action 1, prefix 206.167.100.0/24
: No buffer space available
Aug 10 14:04:41 btn-gw bgpd[5238]: send_rtmsg: action 1, prefix 206.167.101.0/24
: No buffer space available
Aug 10 14:04:41 btn-gw bgpd[5238]: send_rtmsg: action 1, prefix 206.167.102.0/24
: No buffer space available
Aug 10 14:04:41 btn-gw bgpd[5238]: send_rtmsg: action 1, prefix 206.167.103.0/24
: No buffer space available
[...]
Aug 10 15:39:47 btn-gw bgpd[5238]: send_rtmsg: action 1, prefix 192.138.64.0/24:
 No such process
Aug 10 15:39:47 btn-gw bgpd[5238]: send_rtmsg: action 1, prefix 192.108.183.0/24
: No such process
Aug 10 15:39:47 btn-gw bgpd[5238]: send_rtmsg: action 1, prefix 199.113.128.0/19
: No such process
Aug 10 15:39:47 btn-gw bgpd[5238]: send_rtmsg: action 1, prefix 199.113.128.0/18
: No such process
[...]

As far as I can see, the error message comes from the piece of code
where bgpd passes its routes off via an AF_ROUTE socket to the kernel.

The routes I've checked on do seem to be fine at the moment, though.
Not that there's much comfort in that.

For what it's worth, the error message is eeringly familiar to what I
get much more often from Zebra:

Aug 10 07:51:34 border zebra[416]: write : No such process (3)
Aug 10 07:52:08 border last message repeated 22 times
[...]
Aug 10 16:00:27 border zebra[416]: write : No buffer space available (55)
Aug 10 16:01:03 border last message repeated 145 times
Aug 10 16:02:59 border last message repeated 145 times

Ideas, anyone?


-- 

Arvid