bitcoind crashes when logging time data #589

I'm running version 0.4.0 (observed in 0.3.24 too) on a couple Linux servers.  Every 2-3 weeks, the daemon crashes and I have to restart it.  In every instance, the last entry in debug.log is one like the following:

Added time data, samples 28503, offset +0 (+0 minutes)
-747598944  -747551416  -747428248 ... -201976  -1940

The daemon crashes while it's printing the list of time samples.  In the above example, the median offset was "+0" but "-1940" was the last sample printed.

I haven't noticed any other patterns.  Is there any other information I can provide to help narrow down the problem?

Original posted on the forum

"The daemon crashes" -- does it crash with a segfault or does it stop responding to JSON-RPC requests?

What is it busy doing-- responding to getwork requests? Handling users' wallets? Is it under heavy JSON-RPC load?

Can you compile -g -ggdb and run under the gdb debugger?

Can you compile -DDEBUG_LOCKORDER and then run? Previous hangs were due to deadlocks between threads...

Oh, and what variation of Linux?

does it crash with a segfault or does it stop responding to JSON-RPC requests?

As far as I can tell, there's no segfault. The process simply exits. That made me think the OOM killer had killed the daemon, but kernel logs show no OOM killer activity in the last 2 years.

What is it busy doing-- responding to getwork requests? Handling users' wallets? Is it under heavy JSON-RPC load?

Other than activity required for staying connected as a peer, the daemon is completely idle. I manually execute an API request once every few days, but otherwise it receives no API activity.

Can you compile -g -ggdb and run under the gdb debugger?

Can you compile -DDEBUG_LOCKORDER and then run? Previous hangs were due to deadlocks between threads...

So far I've been unable to compile Bitcoin from source. I'll try again now that QT is merged. I'll post back with further details.

Oh, and what variation of Linux?

Debian 6.0.3

Any updates on this? Anyone been able to reproduce this?

I haven't had this problem since 1c4aab926e0b25e0c896d4703d858eda1eb14ea5

Ok, great to hear that!

I'll close this issue then, you can post here or open a new one if you happen to have the problem again.