[ClusterLabs] pacemaker after upgrade from wheezy to jessie

[Kristoffer Grönlund]

Toni Tschampke <tt at halle.it> writes:

> Hi,
>
> we just upgraded our nodes from wheezy 7.11 (pacemaker 1.1.7) to jessie 
> (pacemaker 1.1.15, corosync 2.3.6).
> During the upgrade pacemaker was removed (rc) and reinstalled after from 
> jessie-backports, same for crmsh.

You'll want to switch your validate-with schema to a newer schema, and most
likely there will be one or two things that don't validate
anymore. There is the "crm configure upgrade" command, but if crmsh is
having problems you can call cibadmin directly:

    cibadmin --upgrade --force

Going from 1.1.7 to 1.1.15 is quite a big jump, so there is a lot that
could go wrong..

Your configuration looks fine from a first glance, the reason you're
getting XML mixed in is because of the missing schema: crmsh can't be
sure that it translated the XML to line syntax correctly, so falls back
to showing the XML. That should all fix itself by changing the
validate-with attribute on the <cib> root tag to a newer version.

I'm guessing that the errors you are getting when connecting the second
node are due to the missing schema, it's hard to tell based on the log
snippet attached though.

-- 
// Kristoffer Grönlund
// kgronlund at suse.com