Re: at-ws.ten-34.net instabilities

[Felix Kugler <]

Michael,

> For your information: Over the last days we have experienced several times
> that mrouted dies (core dumped) on the at-ws.ten-34.net. Also, at times
> tunnels appear to be down for some time, without obvious reason. The
> mrouted version is currently 3.9-beta2. The whole set-up is exactly the
> same as on the other TEN-34 mrouteds, which have no problem (knock on
> wood). We will try upgrading to the new beta3 after the IETF. Any other
> ideas: Please let us know ;-)

that's probably a good move, we too experienced core dumps of
mrouted3.9b2, Simon was even able to find one of the bugs in the code
which caused the crash. According to Bill Fenner this has been fixed
and we indeed did not get any crashes of mrouted since the upgrade to
3.9b3.

> PS: Yesterday the CH-WS went down hard (no telnet possible anymore) twice,
> too, which is probably not related to mrouted, but will have had effects on
> the mbone as well. 

I'm afraid it is. Unfortunatly, we notice this behaviour quite
frequently, especially yesterday and this morning on all our three
mrouted's in Geneva, Zuerich, and Lausanne. Often the machines seem to
freeze for 20..30 seconds and come to live again, sometimes they simply
hang forever. The freeze periods begin after short periods of very
high CPU load (nice to watch with "iostat -5"). Traffic seems to peak
too during these periods. 

Anybody else getting these problems ?

Felix

-----------------------------------------------------------------------------
Felix Kugler                                     | email: kugler@localhost 
SWITCH - The Swiss Academic and Research Network | Phone: +41 1 268 15 30 
Smail:  Limmatquai 138, 8001 Zurich, Switzerland | FAX:   +41 1 268 15 68