Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
Just in case you're using LVM as well - I had issues with the heartbeat resource agents for a crm style configuration. The RA didn't properly deactivate the volume group, and it gave me the same error message you are reporting. I wound up modifying the RA to fix it here. Justin -----drbd-user-bounces at lists.linbit.com wrote: ----- To: "Nikhil Kulkarni" <Nikhil_Kulkarni at symantec.com>, <drbd-user at lists.linbit.com> From: "Nikhil Kulkarni" <nikhil_kulkarni at symantec.com> Sent by: drbd-user-bounces at lists.linbit.com Date: 07/10/2008 07:33PM Subject: RE: [DRBD-user] Device is held by someone error. I think I might have found the problem. It looks like heartbeat is not shutting down the nfs service cleanly, it still leaves certain threads open. I did a kill -9 on the nfsd daemon and then it looks like the server is still up. I’m going to do some more digging into this and will keep you all posted. Thanks, Nikhil From: drbd-user-bounces at linbit.com [mailto:drbd-user-bounces at linbit.com] On Behalf Of Nikhil Kulkarni Sent: Thursday, July 10, 2008 3:44 PM To: drbd-user at linbit.com Subject: [DRBD-user] Device is held by someone error. Hi, I am getting the following error on a CentOS + DRBD + HeartBeat setup: When I stop the heartbeat service on the primary, for some reason it cannot change the state of drbd using the drbdadm command and I get the below error in the log file: State change failed: (-12) Device is held open by someone Manually running the “drbdadm secondary r0” command does not work either. The file system is not mounted, running lsof does not show any processes accessing the drbd device and Fuser –mv returns nothing. I also ran “grep drbd /proc/mounts” and that too returned nothing, which tells me that the file system is un-mounted cleanly. I’m going to install strace and see if that provides me with any more information. After this the system just reboots. I googled around, went through a whole bunch of mailing lists and did not see a reason as to why this was happening and what can be done to avoid it. Is this just a bug with drbd or heartbeat? Has anyone else seen and fixed this issue? When I googled, I did come across people having a similar problem, but there was no mention of how it was resolved. Will upgrading to the latest drbd release solve it? My setup: CentOS 5.1 DRBD : 8.0.11-1 Heartbeat : 2.1.3-3 I did not want to paste the drbd.conf file since the post would be pretty long, but if it is needed then please let me know and I’ll post that information as well. I’m really running out of ideas, any help is greatly appreciated. If anyone needs any more information from my end, then please let me know and I’ll be happy to provide that too. Thanks, Nikhil _______________________________________________ drbd-user mailing list drbd-user at lists.linbit.com http://lists.linbit.com/mailman/listinfo/drbd-user