Skip to contentSkip to conversation

BIND hangs with 2.6.14

DISCUSSION Stefan Schmidt,,, 2 participants 2 messages Oct 31, 2005 3 messages not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

#3 lore raw
Quoted 3 lines · you

Unfortunately, the machine does quite a bit of other work apart from BIND, so unless somebody can reproduce this on another machine, it will be a bit difficult.

I saw the same error with 2.6.14-rc5 and filed bug #5505 against it. I was also able to reproduce the behaviour using the same kernel and same library/compiler versions on another machine using the same set of zones and tcpreplay to simulate the query load. Our machine gets around 10 queries per second at ~40k slave zones. BIND version is 9.3.2v1 which runs smoothly on our in-production nameserver with kernel 2.6.13.4 now.

netdev and the usual suspects are aware of the problem.

Stefan

#4 lore raw
Quoted 6 lines · Stefan Schmidt,,,
Quoted 3 lines · you

Unfortunately, the machine does quite a bit of other work apart from BIND, so unless somebody can reproduce this on another machine, it will be a bit difficult.

I saw the same error with 2.6.14-rc5 and filed bug #5505 against it. I

I'm suprised 2.6.14 went out the door without this being resolved. I'm glad I spotted this thread in time to avoid upgrading my servers! I guess it was reported on the cusp of the release though.

Quoted 7 lines

was also able to reproduce the behaviour using the same kernel and same library/compiler versions on another machine using the same set of zones and tcpreplay to simulate the query load. Our machine gets around 10 queries per second at ~40k slave zones. BIND version is 9.3.2v1 which runs smoothly on our in-production nameserver with kernel 2.6.13.4 now.

netdev and the usual suspects are aware of the problem.

I guess I'll follow this on bugzilla.

Andrew Walrond