[DRBD-user] Notify admin for online verification

[Jeffrey Froman]

Hello All,

I am still trying to enable the out-of-sync handler,
such that online device verification executed from a
cron job will notify an administrator if any blocks are
found to be out of sync.

So far I have been unsuccessful: online device
verification, whether run from the command line
manually, or from cron, does not trigger my handler.

I have followed the example described in
http://git.drbd.org/?p=drbd-8.2.git;a=blob;f=scripts/drbd.conf
My full configuration is reproduced below, along with my
handler script and log output from "drbdadm verify".

I have ensured that my handler script works as expected
when invoked directly on the command line. I have also
tried using a simple inline script for my out-of-sync
handler (e.g., "echo test | mail root"), with the same
results: my handler is never invoked.

So my questions at this point are:

1. Am I making any obvious mistakes or omissions in my
configuration?

2. Is there any documentation for the out-of-sync handler
beyond the example configuration cited above?

3. Has anyone else successfully enabled this handler?

4. Is there a way to increase log output regarding the
invocation of this handler, so that I can better determine
exactly why it is not running?

I am using CentOS-4.6, drbd-8.2.5, kernel 2.6.9-67.0.7.EL.
My configuration is as follows:
---------------------------------------------------
global { usage-count yes; }
common {
    startup {
        wfc-timeout  20;
        degr-wfc-timeout 20;
    }
}
resource spare {
     protocol C;
     syncer {
         rate 100M;
         verify-alg sha1;
     }
     net {
         cram-hmac-alg sha1;
         shared-secret "foo";
     }
     handlers {
         out-of-sync "/opt/drbd/handlers/out-of-sync.sh $DRBD_RESOURCE";
     }
     disk {
         size 1703936K;
         on-io-error pass_on;
     }
     on host1 {
         device     /dev/drbd0;
         disk       /dev/VolGroup00/LogVol09;
         address    172.21.103.1:7788;
         meta-disk  internal;
     }
     on host2 {
       device     /dev/drbd0;
       disk       /dev/VolGroup00/LogVol07;
       address    172.21.103.2:7788;
       meta-disk  internal;
     }
}
 -------------------------------------------------------------------

Running the command "drbdadm verify spare" on host1 produces
this output in /var/log/messages (is the "helper command"
supposed to be my handler?):

 drbd0: conn( Connected -> VerifyS )
 drbd0: Out of sync: start=0, size=16 (sectors)
 drbd0: Out of sync: start=840, size=24 (sectors)
 drbd0: Out of sync: start=4952, size=8 (sectors)
 drbd0: Out of sync: start=5000, size=96 (sectors)
 drbd0: Out of sync: start=70616, size=8 (sectors)
 drbd0: Online verify  done (total 39 sec; paused 0 sec; 43688K/sec)     
 drbd0: Online verify found 19 4k block out of sync!
 drbd0: helper command: /sbin/drbdadm out-of-sync
 drbd0: Writing the whole bitmap.
 drbd0: writing of bitmap took 2 jiffies
 drbd0: 76 KB (19 bits) marked out-of-sync by on disk bit-map.
 drbd0: conn( VerifyS -> Connected )
 _______________________________________________

My handler script is as follows:

 #!/bin/bash

 admin_address="root"
 subject="Alert: Out-of-sync drbd resource found."
 alert="
    Found out-of-sync drbd resource $1 on $HOSTNAME.
    To re-synchronize, run:\n
    \n
    # drbdadm disconnect $1\n
    # drbdadm connect $1\n
 "

 /bin/echo -e "$alert" | /bin/mail -s "$subject" "$admin_address"
_______________________________________________


Thank you,
Jeffrey