Mailing List Archive

DRBD Filesystem Pacemaker Resources Stopped
Thank you both for your responses. Interesting. I couldn't find them until today. I guess I wasn't looking well enough in the April archive. I searched using Google and found them on; then went back to the mailing list archive.

So, I was able to find the issues myself and am excited to announce that I had the cluster successfully working as of Friday, April 13 at 1:30pm PST. Unfortunately, I wasn't taking notes very well; but okay, I think enough to know what I would need to look out for on a future build.
From what I did note, I found that I was missing the closing right curly brace for the DRBD resource (nfs.res).
I commented out the Startup section entries that I added.

@Andreas; Thank you for sharing your knowledge on my not needing the Startup entries that I added. It is a nice validation. I was not sure. And also about letting me know that I need to put Pacemaker into maintenance mode. Funny, I just read about that the other day too. I'm sure I have a lot more to learn and I'm eager to as well.

I also found that my CIB was broken; but after I started adding the NFS Export resources (This was after I found out about the DRBD resource). I missed an ending quote around one of the exportfs resource directories (Vol01).

When creating the resources, I kept receiving errors about he default timeout being set to 20 seconds and the LinBit NFS tutorial (which I am so grateful for) was calling for the monitor interval to be set at 30 seconds on many of them.
So, I changed the default timeout to 40 seconds and that seems to take care of the problems I continued to have with the resources.

Another problem happened from inexperience. I had created my exports in the /etc/exports file, not realizing that the exportfs OCF resources were already creating the NFS exports. I realized that from running the mounts command.

Now, I have another issue while attempting to perform a test failover. I'll send in a separate email about that.

I appreciate all of the help.


drbd-user mailing list
DRBD Filesystem Pacemaker Resources Stopped [ In reply to ]
BTW; @Andreas, Thanks for pointing out that I'm not using STONITH and DRBD resource level fencing. I have STONITH off at the moment since that is what the LinBit document called for and because I don't know enough about it at the moment to put it into play. But, I have had it on my mind to enable it.
I'm reading up on STONITH via what I can find on ClusterLabs site and whatever else comes up on a Google search.
If you have any suggests, I am grateful for whatever you can pass on.
I'll also be sure to read up on fencing. I'm wondering if that is related to what I'm currently facing after attempting a fail-over. (My next email to the mailing list...)

drbd-user mailing list
DRBD Filesystem Pacemaker Resources Stopped [ In reply to ]
Never mind about my new issue after fail-over. I found more of my mistakes. Ironing out a lot of them it seems.
Searching syslog I found the Heartbeat error about messages being too long. I've now turned on Compression. And I also did not specify bcast on the ds02 node (my bad for not copying the file over and thinking that I could do a manual copy). All appears to be working now. Time to setup DRBD resource fencing, learn about implementing STONITH, then, another fail-over test to fail back to ds01.

Good things come to those who persevere. Right? -Yeah

drbd-user mailing list