On 2009-05-01T16:38:50, [email protected] wrote: > Goal: I'm trying to get a 2 node Active/Passive cluster working with > DRBD replication, an ext3 FS on top of DRBD and a virtual IP. I want the > active node to have a mounted FS that I can serve requests from using > ProFTPD or another FTP daemon. If the active node fails I want the > cluster to migrate all 4 resources (DRBD, FS, ProFTPD, Virtual IP) > across to the other node. I don't have any STONITH devices at the > moment.
You really should configure STONITH. The manual which you got with SLE11 HA is quite insistent on that ;-) > Problem1 - Using DRBD OCF RA: I wanted to use the latest and greatest > for the approaches, so tried the DRBD OCF RA following this howto: > http://www.clusterlabs.org/wiki/DRBD_HowTo_1.0 . The configuration works > and I can manually migrate resources but if I just reboot the node that > has the drbd resource on it I see the resource gets migrated to the > other node for about 2 seconds then is stopped: Yes, this happens because your replica loses quorum, you need to set the cluster property no-quorum-policy to "ignore" for two node clusters. Andrew has changed this in pacemaker 1.0 instead of continuing to pretend that 2 node clusters have quorum - people now have to make an explicit choice. > Problem 2 - Using heartbeat1 style drbddisk RA: As I read that Linbit > devs recommend to use the old heartbeat-1 style drbddisk RA I decided to > go that route. I found the haresources2cib.py script, used it and got > the XML into the CIB. crm_mon showed lots and lots of errors. The XML > generated is: Why bother with creating the configuration using the haresources2cib script? The "crm" shell is quite excellent, and the documentation contains all the bits you need to setup drbd on SLE11 HA. (I think haresources2cib generates SLES10 style XML, which the tools can consume, but only in legacy mode - ie, you can't mix it with a cluster created on SLE11 natively.) You'll find the crm shell is about as terse as the "haresources" style, but much, much more powerful. > So after googling I find you have to replace the CIB rather than create > it even though I had already run cibadmin -E --force . Now it goes in > but the following happens: > > gihub1:~ # crm_mon -1 Yes, seems the tool tries to validate the old style CIB with the new schema. Not good. Please, try the documentation. (On SLE11 HA, it's in the sle_ha_en_pdf package, if I remember the name correctly.) I could copy&paste the relevant section from the manual here, but I want you to read it by yourself, which is probably more helpful in the long run ;-) If there are remaining questions, please feel free to ask. Regards, Lars -- SuSE Labs, OPS Engineering, Novell, Inc. SUSE LINUX Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg) "Experience is the name everyone gives to their mistakes." -- Oscar Wilde _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
