[DRBD-user] DRBD 0.8rc2 deadlock "WFBitMap" ?
philipp.reisner at linbit.com
Tue Jan 23 16:21:18 CET 2007
Am Montag, 22. Januar 2007 19:09 schrieb Ralf Schenk:
> Philipp Reisner schrieb:
> > drbdadm down all ; drbdadm up all
> That doesn't work: Error is:
> Child process does not terminate!
> No response from the DRBD driver! Is the module loaded ?
> (Of course the module is loaded)
> > But, please post your excerpts from the kernel logs (of both machines)
> > of the incident.
> > And tell us if the clocks of the machines are in sync (with NTP), or
> > what time difference you have there.
> Both nodes dates are in sync with ntp.
> Ok what I did now: created the meta-data completely from scratch on both
> nodes (drbdadm create-md vm0[1,2]). I started drbd (/etc/init.d/drbd
> start). I set each of the two nodes primary for one drbd device (drbdadm
> primary -o vm0[1,2]). I expected drbd to start a full resync...
So far quite interesting. But to understand if there is really a bug
May I ask you to send the log from the other machine as well ?
I need it of __BOTH__ machine to understand what is going on there!
Hmm, you can reproduce the issue. Maybe you can send me enough information
so that I can reproduce it here as well.
1) Your drbd.conf
2) The drbd-meta-data of both devices on hosta
3) The drbd-meta-data of both devices on hostb (I suppose it has that name!)
This is the way to create meta-data-dumps:
drbdadm dump-md all > meta-data-of-hosta.txt
PS: Please attach the log snipplet as mime attachment, (or try to
not break long lines with you e-mail client). The same
holds true for the meta-data-dumps.
: Dipl-Ing Philipp Reisner Tel +43-1-8178292-50 :
: LINBIT Information Technologies GmbH Fax +43-1-8178292-82 :
: Vivenotgasse 48, 1120 Vienna, Austria http://www.linbit.com :
More information about the drbd-user