see also this nice & actual howto:
http://clusterlabs.org/mw/DRBD_HowTo_1.0
Yes, that was written by me.
But - pretty much all I did was replace the xml snippets with working
syntax for pacemaker 1.0
The rest of it is almost completely identical with Lars Marowsky-Bree's
doc for older versions: http://wiki.linux-ha.org/DRBD/HowTov2 (I even
said so in the doc).
exactly this howto I was referencing I had in my mind. The developers say,
that v8 is not 100% tested and OCF RA does not work with Master/Master. My
questions were:
- is it working?
Dominik seems to think so, based on his testing, and I've no reason to doubt him
Wait wait wait. I guess Andrew misunderstood something here.
I _never_ said master/master works and I _never_ tested that with any RA
in any version of pacemaker.
What I did say was that drbd, the ocf ra and pacemaker do support
multiple drbd instances, which can all run in master/slave mode. That is
to say /dev/drbd0 and /dev/drbd1 can be managed by one cluster but as
two resources.
- Master/Master with drbdlinks or with OCF RA?
I think he's using the OCF RA.
I think I tried out drbdlinks _once_, but never looked any further.
Regarding v8 features:
I tried getting dopd onboard in my test system last week but again I
could not get results I liked. I guess it needs some playing with
timeouts, but I can't say for sure.
What I also tested was the verify feature of v8. The RA does work while
the drbd device is verified. Looking at the drbd users-guide, this is
not very surprising:
"Any applications using the device at that time can continue to do so
unimpeded, and you may also switch resource roles at will." Quote from
http://www.drbd.org/users-guide/s-use-online-verify.html
Regards
Dominik
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
