Update time. It's been a week now since I decided to start rolling back kernels, since then my FC5 boxes running 2.6.16-1.2108_FC4 have been fine. But yet another of my 2.6.16-1.2122_FC5 boxes has locked up with the same symptoms, on top of the previously reported behavior of the most basic functions (lstat for example) taking thirty seconds or more to complete and time running backwards and/or irregularly, I've now seen an odd ping report : # ping 192.168.69.9 ..<I had to wait about 15-20 seconds before this returned anything>.. PING 192.168.69.9 (192.168.69.9) 56(84) bytes of data. 64 bytes from 192.168.69.9: icmp_seq=1 ttl=64 time=0.058 ms ..<the first packet took another ~30 seconds>.. --- 192.168.69.9 ping statistics --- 1 packets transmitted, 1 received, 0% packet loss, time 0ms rtt min/avg/max/mdev = 0.058/0.058/0.058/0.000 ms And the report states everything is normal which I wasn't expecting. So I'm reasonably confident this nasty bug which is biting my production servers is related to a recent FC5 kernel. I'll continue to roll forward from my FC4 kernels until I hit my typhoid Mary. This could be a long process though. I'll take a look at forcing a core dump during the next occurrence as well. On Fri, 2006-06-09 at 12:52 +0200, Bart Couvreur wrote: > 2006/6/9, Naoki <naoki@xxxxxxxxxxxxxxxxx>: > > > > > Jep I had this last night, while running yum: I started it at 11:45 pm > > > and when I looked at my box this morning yum had stalled and the time > > > was 0:45 am (the real time was 8:30 am). And ntpd is also running. > > > > > > I'm suspecting this has something to do with the kernel, but not sure. > > > (running 2.6.16-1.2122_FC5). I've looked through all the logs, but it > > > doesn't mention a thing. > > > > > > I hope this gets solved soon, has been quite irritating, > > > Bart > > > > Irritating isn't the word when it's happening to production servers ;) > > Yeah sure :) > > > I've been through a bunch of patches for kernel, glibc, and ntp (client > > + server ) over the last few weeks but so far nothing has helped. I'm > > hoping that reverting to the FC4 kernel will at least prove where the > > problem lies. After that though I'll probably not be able to diagnose > > much more. > > > > If others are noticing the problem though I'll keep you posted. > > If you could that, would be great. > > > By the way, are you using ResiserFS or ext3 ? > > It's ext3 here. > > Bart >
