Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
Lars Ellenberg wrote: >/ 2006-02-21 09:28:14 +0100 > \ Michael Paesold: >> Just some analysis from myself: >> >> Michael Paesold wrote: >> >Logs from primary/sync source: (xen03) > ^^^^^^^ > > this is not "plain linux", but Xen master virtualization domain, > or what are you up to? This is plain linux. It should eventually become Xen virtualized, but right now, the server has not seen a single bit of Xen. ;-) > well. we could not start the thread (from the output below it is the > "worker", that failed to start), so the task pointer is NULL. > > guess what, if the worker is not there, who does the work? > thus: stalled. It's not only that, it *seems* this finally resulted in a kernel panic (I could send you the logs), because after that the issue propagated even further and somehow managed to cause major havoc :-). Could this error, i.e. a failing kernel_thread be handled somewhat more gracefully? Is there any way to back out for the drbd kernel module and say: huh, I give up... > why the thread could not be started? > I don't know. Do you know where I could find what "-4" means as the result to kernel_thread? Perhaps this is broken kernel, but then, it's the stock CentOS-4.2 (rebuilt RHEL 4 update 2) kernel. Best Regards, Michael Paesold