[DRBD-user] Failing to migrate two DRBD nodes

Marc Richter drbd at zoosau.de
Wed Jan 12 16:17:05 CET 2011

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

Hi Dan,

when I'm using a "fresh" SyncTarget and connect it to the SyncSource,
then the content of /proc/drbd is set to "WFBitMapS" on Source and
"WFBitMapT" on Target and I cannot access the Filesystem on the Source
anymore because of the high I/O.

Syncer Rate is set to 40M currently, but this is the same with 20M or
even lower.
One has told me once, this rate determines the rate for a already
running cluster only and doen't affect the initial sync at all.
Don't know if this is true, but seems to fit what I'm seeing here ..

Am 12.01.2011 16:02, schrieb Dan Barker:
> I'm not sure why you say "makes the devices quite unusable", I'm wondering
> why. Is your syncer speed set way too high? If you only allow syncer 20% of
> your bandwidth, the impact should be nearly nothing. It might take more
> clock time, but you wouldn't be in such a hurry.
> Dan
> -----Original Message-----
> From: drbd-user-bounces at lists.linbit.com
> [mailto:drbd-user-bounces at lists.linbit.com] On Behalf Of Marc Richter
> Sent: Wednesday, January 12, 2011 7:26 AM
> To: drbd-user at lists.linbit.com
> Subject: [DRBD-user] Failing to migrate two DRBD nodes
> Hi There.
> I'm still failing in replacing a HA - node and home someone may help me.
> I'm trying the following:
> I have two nodes which serve as a HA - NAS and are connected by DRBD. We
> have bought new Hardware and installed a new version of the Linux
> Distribution onto this new devices. Since the complete initial sync takes a
> long time and makes the devices quite unusable for 5 hours, I'm planing to
> do the following:
> 1) Remove the secondary node from the cluster by issuing "drbdadm disconnect
> r0".
> 2) Connecting the first new node to this removed node and have the initial
> re sync done without affecting the live node.
> 3) When they have synced, disconnect these two nodes and re-connect the new
> one to the still productive old one, to have the changes synced, which
> occurred during the last few hours since the initial rebuild was running.
> 4) When they have synced, remove the old one from the cluster and have the
> now synced new one serving all requests.
> 5) Connecting the second old node to the second new one and again let them
> do the initial syncing.
> 6) After this is done, reconnect the second new node to the first.
> 7) I'm done with a very short downtime.
> I previously set up 4 virtual machines (2 old ones, 2 new ones) which used
> _exactly_ the same Versions of the OS, which the real nodes are running on
> and copied the Configurations for drbd exactly from the real nodes (changed
> only the IPs).
> This way I could successfully finalize and simulate this plan. All went
> well, so far.
> Now I'm trying the same with the real nodes, but I'm stuck with point 3).
> I've successfully synced the secondary old node with the first new one.
> But when I try to connect this new node to the second (productive) old node,
> I get the following in the Logs: "Unrelated data, aborting".
> I googled for this, but don't understand what has/is happening here:
> I extracted the UUID from the secondary old node (SyncSource) by issuing
> "drbdadm show-gi r0" and got
> "1F583294AF81AF78:0000000000000000:F2997C7C2F263DF4:0F765342B3D081E2:1:1:0:0
> :0:0:0".
> This is exactly the same string I get when I issue the same command on the
> SyncTarget (first new node I connected to this old one).
> When I issue this on the currently productive old one I'm trying to connect
> to, I get
> "C37BCA37822F0D6D:0F765342B3D081E3:2DD3D7AEBE3F7A75:8C5762D1E5ABE255:1:1:1:0
> :0:0"
> As far as I understand all of this, drbd rejects to connect the two nodes,
> because the UUIDs (the part
> "1F583294AF81AF78:0000000000000000:F2997C7C2F263DF4:0F765342B3D081E2"
> without the trailing state Bits) differ (correct?).
> But how can this be? The sync source I used was connected to the other node
> and the whole thing worked perfectly for a long time.
> And what can I do now to connect these two nodes?
> Can / should I change a part of the UUID - String on any of the nodes or
> such? And if so, how?
> Thanks very much in advance.
> Best regards,
> Marc
> _______________________________________________
> drbd-user mailing list
> drbd-user at lists.linbit.com
> http://lists.linbit.com/mailman/listinfo/drbd-user
> _______________________________________________
> drbd-user mailing list
> drbd-user at lists.linbit.com
> http://lists.linbit.com/mailman/listinfo/drbd-user

More information about the drbd-user mailing list