[DRBD-user] Reconnecting after WFBitMapS/WFBitMapT

Ben Beuchler insyte at gmail.com
Fri Jan 29 04:23:32 CET 2010

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


I just tried to connect up a remote node after performing Truck Based
Replication of a 525G volume.  I changed the IPs in /etc/drbd.conf and
ran "drbdadm --stacked adjust all" on the primary.  The resource
connected immediately, both nodes moved to WFBitMap[ST] and stayed
there, the load shot way up on the primary, and the mounted partition
became unwritable.  While I could read it locally on the primary,
boxes that had the volume mounted over NFS were unable to read or
write.

When I noticed this, I ran "drbdadm --stacked disconnect all" and,
after a minute or so, the resource disconnected and everything became
usable again.

The link between the sites is capped at 1 megabit and I had forgotten
to change the syncer rate down from the initial Gb speeds.  Could
trying to sync at 900 megabit over a 1 megabit connection have caused
the problem?

If so, now that I've reduced the sync rate to 100K (or 800 kilobit),
will I still get stuck with a disk that's unusable for quite some time
after re-connecting the remote resource?

If it matters, the connection between the two sites is via an OpenVPN tunnel.

Thanks!

-Ben



More information about the drbd-user mailing list