[DRBD-user] drbd newb help

[Kelsey Cummings]

I've just finished setting up drbd replication between for two volumes
between two hosts ( sda -> sda, sdb <- sdb ) which was easy enough to get
setup.  Dual Xeon servers with 3ware 9xxx series controllers, intel gig
nics for the sync interface, running centos 5, x86-64.

b, which is currently being synced to from a, started spewing this out to
syslog:

Sep  6 21:29:59 b kernel: drbd1: ASSERT( atomic_read(&mdev->local_cnt)>=0 ) in /home/buildsvn/rpmbuild/BUILD/drbd-kmod-8.0.4/_kmod_build_/drbd/drbd_int.h:1769
Sep  6 21:29:59 b kernel: drbd1: ASSERT( atomic_read(&mdev->local_cnt) ) in /home/buildsvn/rpmbuild/BUILD/drbd-kmod-8.0.4/_kmod_build_/drbd/drbd_actlog.c:826
Sep  6 21:29:59 b kernel: drbd1: ASSERT( atomic_read(&mdev->local_cnt)>=0 ) in /home/buildsvn/rpmbuild/BUILD/drbd-kmod-8.0.4/_kmod_build_/drbd/drbd_int.h:1769
Sep  6 21:29:59 b last message repeated 140 times

Any clues?  Anything else needed to help troubleshoot?

my drbd.conf

common {
    syncer {
        rate             500M;
    }
}

resource r0 {
    protocol               C;
    on b.store. {
        device           /dev/drbd0;
        disk             /dev/sda;
        address          10.2.0.1:7788;
        meta-disk        internal;
    }
    on a.store. {
        device           /dev/drbd0;
        disk             /dev/sda;
        address          10.2.0.0:7788;
        meta-disk        internal;
    }
    net {
        cram-hmac-alg    sha1;
        shared-secret    foo;
        max-buffers      4096;
        max-epoch-size   1024;
    }
    disk {
        on-io-error      detach;
    }
    syncer {
        rate             100M;
    }
}

resource r1 {
    protocol               C;
    on b.store. {
        device           /dev/drbd1;
        disk             /dev/sdb;
        address          10.2.0.1:7789;
        meta-disk        internal;
    }
    on a.store. {
        device           /dev/drbd1;
        disk             /dev/sdb;
        address          10.2.0.0:7789;
        meta-disk        internal;
    }
    net {
        cram-hmac-alg    sha1;
        shared-secret    foo;
    }
    disk {
        on-io-error      detach;
    }
    syncer {
        rate             100M;
    }
}

I've applied the standard 3ware tuning values, as well as bumped up the
nr_requests to 1024 per some info I dug up on google.

echo 64 > /sys/block/sda/queue/max_sectors_kb
/sbin/blockdev --setra 16384 /dev/sda
echo 1024 > /sys/block/sda/queue/nr_requests
echo deadline > /sys/block/sda/queue/scheduler


-- 
Kelsey Cummings - kgc at corp.sonic.net      sonic.net, inc.
System Architect                          2260 Apollo Way
707.522.1000                              Santa Rosa, CA 95407