[DRBD-user] corrupt AL transactions

Andreas Poenicke admin at tfp.uni-karlsruhe.de
Thu Mar 19 15:40:30 CET 2015

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

Dear all,

in the preparation of a system upgrade, we ran a number of up-/
downgrade tests in a (kvm) virtualized environment.

During these tests we encountered the error message

"1 corrupt AL transactions found"

which seems to be sticky, meaning getting it every time we
do a "drbd up" on the concerned DRBD resource.

Surprisingly we couldn't find any information about this message
except the original source. It would be nice if anybody could
point us to some documentation or help us with the questions:

1.) What does this message mean?
2.) Is it serious?
3.) Is there any way to "repair" the system?

The steps which led to the problem where

  1.) Two nodes, running Debian Jessie, DRBD 8.4.3
  2.) Downgrading DRBD 8.4 to 8.3
     Reinstalling of the nodes with Debian Squeeze,
     kernel 3.2.63-2+deb7u2~bpo6+1,    DRBD 8.3.11
  3.) Everything seemed fine
  4.) Upgrading first node again to Debian Jessie
     - drbd down, shutdown, unplugging device
     - installation, shutdown, plugging in device, boot
  5.) On" drbd up" the error message appeared.
  6.) Second node (still drbd 8.3) shows no error
  7.) Upgrade of second node, same error on second node
  The error occurs at only one of two drbd-resources, both handled the
  same way.
  Note that we did not encounter any problems with the system other than
  this error message being shown at drbdadm up. Other than that, DRBD
  appears to work properly.
While the way the problems have been generated is unusual, we are
concerned how to handle the message if it occurs in a production environment.
Especially since the  only search hit for the message is the source  	
(drbdmeta.c in drbd-8.4 , source code search engine).


More information about the drbd-user mailing list