I have used SLES 11 SP2 High Availbility Extension and configured DRBD, XEN VM’'s etc without issue.
I am now trialing SLES 12 with its High Availability Extensions. Getting used to Hawk is interesting enough (loved pacemaker GUI).
My issue is that I tried to create a DRBD resource (the same as I did on SLES 11 SP2) and it will not start. I also created a Pingd resource and it too will not start.
This was a given in SLES 11 SP2, yet in SLES 12 I cannot understand why its failing.
An error message Hawk shows is
linux3: WARN: passwordless ssh to node(s) linux4 does not work
Permission denied (publickey,keyboard-interactive).
tar: This does not look like a tar archive
tar: Exiting with failure status due to previous errors
Where linux3 and linux4 represent my 2 SLES 12 servers.
I do not recall configuring SSH for HA in SLES 11 SP2 before - it just worked.
I cannot seem to find any documentation on this error or how to set it up for HA use.
I setup SLES 11 SP3 HAE a couple of months ago and as part of that I used
the sleha-init script which sets up password-less SSH access. I used HAWK
in that environment as well for almost everything since that made a since
remotely-accessible UI for the customer for whom I was implementing it:
I found that starting the ms drbd resource brings up drbd on both nodes into slave/secondary state. However, when the cluster tries to promote it to master/primary state, it fails with a message in the log stating that the drbd resources need to be in an ‘uptodate’ state.
I am able to bring the drbd devices on both nodes to primary state without issues manually-outside of the cluster manager.
If I then performes a ‘clean up’ on the ms resource, the ms resource brings the drbd devices to master/primary state on both nodes without error.
I can then perform a demote, promote, and stop successfully on the running ms resource.
I decided to see what would happen if I shutdown the 2 nodes and powered them up again. I was happy to see the the ms resource went into master state without error.
I have successfully set up passwordless SSH communications between my 2 nodes (found good documentation on this using the ssh-kegen command and copying the resulting id_rsa.pub key to the authorized_key file of the other node).
My drbd issue is now resolved (please see my other post on this).
On 03/25/2015 05:34 PM, 6529034 wrote:[color=blue]
I have successfully set up passwordless SSH communications between my 2
nodes (found good documentation on this using the ssh-kegen command and
copying the resulting id_rsa.pub key to the authorized_key file of the
other node).[/color]
Well done, and thank-you for posting back your results. For fun, you may
want to look at the ssh-copy-id command which will do the pub key copying
to the remote authorized_keys file for you, and be sure permissions are
correct, and with current versions (on my openSUSE 13.1 system anyway)
copy multiple keys (if available), but only if not already copied (to
avoid duplicates, in case you’ve copied in the past).
ssh-copy-id user@remote.system.here
–
Good luck.
If you find this post helpful and are logged into the web interface,
show your appreciation and click on the star below…