14
votes
Accepted
How to get DRBD nodes out of Connection State StandAlone (and WFConnection)?
The situation was apparently caused by a case of split-brain.
I had not noticed this because I had only inspected recent journal entries for drbd.service (sudo journalctl -u drbd), but the problem ...
4
votes
Accepted
Why to use fencing when you already have redundant communication links?
The idea is that when your cluster decides it requires to fail-over a node, there must be something wrong with that node.
Fencing, shooting the other node in the head (STONITH) is the best guarantee ...
3
votes
multicast address use in corosync
You don't have to use multicast on corosync.
You may use unicast to do your job.
This can be done by using something like this in /etc/corosync/corosync.conf for a two-member cluster:
compatibility: ...
2
votes
Correct use of Pacemaker's ocf-tester with OCF resource agents
DRBD can exist in many states, and therefore requires an additional master/slave set to be defined in Pacemaker that references the primitive which uses the ocf:libit:drbd resource agent. Resource ...
2
votes
Accepted
crm status : 0 nodes and 0 resources configured
It seems you only have two nodes, you'd want to add this:
quorum {
provider: corosync_votequorum
expected_votes: 2
}
2
votes
crm status : 0 nodes and 0 resources configured
in totem section, you can try to add this line:
rrp_mode: none
If it still has error, try to add more this section:
quorum {
provider: corosync_votequorum
expected_votes: 3
}
1
vote
Accepted
Pacemaker location constraints on resources apparently not obeyed
I've overcome the problem by switching to another syntactic form for the location contraints that requires neither references to #id nor #uname:
location loc_stonith_node1 stonith_node1 -inf: node1
...
1
vote
Accepted
Correct use of Pacemaker's ocf-tester with OCF resource agents
Your error tells:
Beginning tests for /usr/lib/ocf/resource.d/linbit/drbd...
/usr/sbin/ocf-tester: 226: /usr/sbin/ocf-tester: xmllint: not found
It doesn't have xmllint to test the agent, so first ...
Only top scored, non community-wiki answers of a minimum length are eligible
Related Tags
linux-ha × 27heartbeat × 13
pacemaker × 11
high-availability × 10
linux × 7
cluster × 5
failover × 5
corosync × 5
drbd × 4
centos × 2
mysql × 2
heartbeat1 × 2
apache-2.2 × 1
amazon-web-services × 1
debian × 1
nfs × 1
mysql-replication × 1
disaster-recovery × 1
sles × 1
keepalived × 1
architecture × 1
virtual-ip × 1
rhcs × 1
unicast × 1
ucarp × 1