SIOS TECHNOLOGY CORP.

DRBD Recovery Kit Known Issues / Restrictions

Description
In a three-node configuration, if only one node is started after two nodes have been stopped, switchover will fail on the started node.

Perform a switchover after starting up all nodes.
The lkcli command for DRBD does not support export/import.
lkbackup will not backup resource files if there are more than one set of resource files.

lkbackup will not backup DRBD resource files when there are more than one set of resource files (one set is defined as lk#.res and lk#.res.db). If there is only one set of files then they will be properly backed up. The workaround is to manually save all resource files.
DRBD: Unable to change DRBD connection endpoint IPs using lk_chg_value.

While the IP addresses used by some LifeKeeper recovery kits can be updated by using the /opt/LifeKeeper/bin/lk_chg_value script (see, for example, Changing the Data Replication Path (DataKeeper), Changing the IP for the Queue Manager (IBM MQ), and Changing LifeKeeper Configuration Values), this script cannot currently be used to modify IP addresses used by DRBD resources.

Executing commands similar to the following:

# lkstop
# /opt/LifeKeeper/bin/lk_chg_value -o 182.168.0.1 -n 192.168.1.1
# lkstart

will modify the IP addresses used by LifeKeeper communication paths on the system. However, this will not update the IP addresses in DRBD configuration files. Usage of LifeKeeper to update the connection endpoints for a DRBD resource is currently not supported.
DRBD does not fully resync after unextend/reextend

If a DRBD resource is unextended and then extended again a full resync is required. If the full resync is not automatically performed take the DRBD resource out-of-service and then back in-service on the same node where it was in-service. This should cause any outstanding out-of-sync blocks to be replicated.
Creating/extending a DRBD resource will fail when protecting NVMe devices and some other devices where the path exceeds 128 characters.

The following error message is displayed when the creation or extension of a DRBD resource fails:
/etc/drbd.d/lk0.res.tmp:48: Token too long >>>/dev/disk/by-id/nvme...<<<

The problem can occur on an NVMe device or other device that does not have a GUID Partition Table (GPT).

LifeKeeper uses the /dev/disk/by-id/ path as a persistent identifier for devices that do not have a GUID Partition Table (GPT) created. If you protect a device with a by-id path that exceeds 128 characters, the creation and extension of the DRBD resource will fail.

Feedback

Was this helpful?

Yes No
You indicated this topic was not helpful to you ...
Could you please leave a comment telling us why? Thank you!
Thanks for your feedback.

Post your comment on this topic.

Post Comment