|Main Archive Page > Month Archives > drbd-user archives|
>From your email "The solution to my particular problem was an improperly formatted haresources file for Heartbeat. Heartbeat was commanding DRBD to unmount/detach a filesystem when a key service was not yet shut down. Once I cleared that up not only did the messages go away but failover occurs within 2-3 seconds!", it sounds like the issue was occurring in case of switchover the machines.
In my case also, the issue "DRBD device held open" is occurring when I am trying to do failover and stuck in device held open issue. I went through the "/etc/ha.d/haresources" file and found that the "drbddisk::drbd0 LVM::vgdrbd0 Filesystem::/dev/vgdrbd0/mysql::/var/TKLC/elap/drbd/mysql::ext3" is unmounting/detaching a file system.
Are you talking about the unmount/detach related to the one highlighted in yellow above. Please provide some thoughts.
From: Felix Frank [mailto:email@example.com]
Sent: Wednesday, March 13, 2013 3:00 PM
To: Satyajit Paul
Subject: Re: [DRBD-user] State change failed: Device is held open by someone
this is an instace of thread confusion. The haresources answer was not actually directed to you, but instead belongs to an old thread that happens to have the same subject.
The proposed solutions applies to setups with the obsolete heartbeat cluster stack, which you are (hopefully) not using, so this will likely not help your case.
On 03/13/2013 06:32 AM, Satyajit Paul wrote:
> Hello Mwoolfso,
> I tested after changing the data type of open_cnt variable to atomic_t, but it didnot work.
> I am planning to test with your fix. Are you talking about the file "/etc/ha.d/ haresources" that you have done code changes to solve the issue ? Can you please let me know what changes u made and if changes in any files other than "/etc/ha.d/ haresources" are required ?
Please refer to http://www.aricent.com/legal/email_disclaimer.html
for important disclosures regarding this electronic communication.
drbd-user mailing list