Hi there,

I'm trying to configure Pacemaker according to http://library.linode.com/linux-ha/ip-f ... untu-10.04

I've done things pretty much exactly as per the wiki, apart from instead of creating two separate drives for /var/lib/mysql and /srv/www, I've just created a /data drive which I intend to symlink to /var/lib/mysql, /home directories etc.

I entered my slightly modified configuration below and it accepted it, but it doesn't seem to start any services or mount the /data mountpoint.

crm_mon output is below as well.

-- snip -- updated config in lower post now

and "crm_mon" shows:

============
Last updated: Fri Oct 15 23:54:37 2010
Stack: Heartbeat
Current DC: ha1 (f691c9d9-17f1-4ae6-9b84-1405abe2957b) - partition with quorum
Version: 1.0.8-042548a451fce8400660f6031f4da6f0223dd5dd
2 Nodes configured, 1 expected votes
3 Resources configured.
============

Online: [ ha1 ha2 ]

Master/Slave Set: ms_drbd_webfs
Masters: [ ha1 ]
Slaves: [ ha2 ]

Failed actions:
fs_mysql_start_0 (node=ha1, call=17, rc=5, status=complete): not installed
fs_mysql_start_0 (node=ha2, call=11, rc=5, status=complete): not installed

Can anyone please help!?

OK - I've fixed it, I'm not sure how but I erased the config and slowly built it up bit by bit (IP, then DRBD, then mounting filesystems, then apache, then mysql) and it's mostly working now.

Got an issue with mysql not wanting to start though now:

============
Last updated: Sat Oct 16 11:01:08 2010
Stack: Heartbeat
Current DC: ha1 (f691c9d9-17f1-4ae6-9b84-1405abe2957b) - partition with quorum
Version: 1.0.8-042548a451fce8400660f6031f4da6f0223dd5dd
2 Nodes configured, 1 expected votes
2 Resources configured.
============

Online: [ ha1 ha2 ]

 Master/Slave Set: ms_drbd_data
     Masters: [ ha1 ]
     Slaves: [ ha2 ]
 Resource Group: WebServices
     ip1        (ocf::heartbeat:IPaddr2):       Started ha1
     ip1arp     (ocf::heartbeat:SendArp):       Started ha1
     fs_data    (ocf::heartbeat:Filesystem):    Started ha1
     apache2    (lsb:apache2):  Started ha1
     mysql      (ocf::heartbeat:mysql): Started ha1 FAILED

Failed actions:
    mysql_start_0 (node=ha2, call=24, rc=4, status=complete): insufficient privileges
    mysql_start_0 (node=ha1, call=51, rc=4, status=complete): insufficient privileges

It lies though, mysql isn't started on ha1 at all...

Well the errors look pretty self explanatory, mysql_start_0 (node=ha2, call=24, rc=4, status=complete): insufficient privileges

Which I'd assume means either pacemaker doesn't have the permissions to run mysql, or the error is coming from mysql and mysql doesn't have the permissions to access it's own files.

Since apache works it's probably a good idea comparing the two permissions and see if there's any difference

I've managed to fix it now thanks!

The issue was mysql couldn't create it's pid and sock files. As /var/run/ is a tmpfs, the normal mysql upstart script creates /var/run/mysqld/ folder with the correct permissions allowing mysql to place pid/socks. However as I'm using pacemaker to start mysql now, the init script didn't create the directories. A bit hacky, but I've just done the folder creation of /var/run/mysqld/ in the /etc/rc.local

It all seems to be working now, unfortunately I still don't know how I managed to get the filesystem mounting properly otherwise I'd say here for others to use in future!

Thanks all for looking and thanks obs for your advice!