[EMAIL PROTECTED] wrote on 06/19/2007 12:16:34 PM:
> Am Dienstag, 19. Juni 2007 schrieb [EMAIL PROTECTED]:
> > Following various sets of documentation I have found on the web I
> > have configured a two node HA/DRDB/NFS setup. Almost everything
> > works without any issues. Running on SLES 10; here are the conf
> > files:
> >
> > cat /etc/ha.d/ha.cf
> > auto_failback on
> > node ulccnfs01
> > node ulccnfs02
> > ucast eth1 10.1.66.110
>
> Only one connection? I use two!
>
> May be, that you get some strange thinks.
> AFAIK, should work without ping hosts.
It is working alright in this configuration for basic failover
>
>
> cat /etc/ha.d/ha.cf: (for version 2 style)
> debugfile /var/log/ha-debug
> logfile /var/log/ha-log
> logfacility local0
>
> keepalive 2
>
> deadtime 15
> #warntime 10
>
> initdead 25
>
>
> bcast vlan255
> bcast bond1
>
> node sot0000140 sot0000149
>
> auto_failback off
>
> crm yes
> use_logd yes
>
> > cat /etc/ha.d/haresources
> > ulccnfs01 10.1.100.140 drbddisk::drbd-resource-0
> > Filesystem::/dev/drbd0::/images::ext3 nfsserver
>
> My haresources looks like this:
>
> node1 IPaddr:10.19.1.119/vlan255 drbddisk::drbd0 \
> FileSystem::/dev/drbd0::/export::ext3 nfssserver nfslock
>
> Now, I use version 2 style.
>
> drbd.conf:
>
> resource drbd0 {
> protocol C;
> incon-degr-cmd "echo '!DRBD! pri on incon-degr' | wall; sleep 60 ;
> halt -f";
>
> startup { wfc-timeout 10; degr-wfc-timeout 120; }
> disk { on-io-error panic; }
> syncer { rate 700000K; group 1; }
> net { timeout 30; connect-int 10; ping-int 10;
> max-buffers 32; max-epoch-size 2048; }
>
> on sot0000140 {
> device /dev/drbd0;
> disk /dev/cciss/c0d0p5;
> address 1.1.1.1:7788;
> meta-disk internal;
> }
>
> on sot0000149 {
> device /dev/drbd0;
> disk /dev/cciss/c0d0p5;
> address 1.1.1.2:7788;
> meta-disk internal;
> }
> }
>
> > I am using a a linked /var/lib/nfs which is linked on the DRBD
> > device so that there is no hiccup on the servers mounting the
> > share. I am wondering if a sleep command in the haresources might
> > give DRBD enough time to mount its partition, if that is the
> > problem, or anything else that may help. The other concern is that
> > it states it is giving up the resources but ulccnfs02 never takes
> > over from this point. Any and all help will be appreciated.
>
> We use SLES9 without linked /var/lib/nfs to the drbd device. we leave
> it original SuSE.
Our nfs mounts go to a webserver cluster. AFAIK if you don't link them you
end up having to refresh the mounts on the servers or you end up with
stale mounts. I am willing to go a different route if there is another
better way.
>
> > The second issue is that after rebooting DBRD ends up in
> > Primary/Uknown and Secondary/Unknown. I know that this isn't the
> > DRBD list but I thought someone here might be able to give some
> > advice.
>
> DRBD NFS works from the console (without heartbeat)?
Yes it works fine. It works fine for the initial failover as well. In fact
I can get it to fail back forth. The only time it doesn't work is failback
from a reboot.
>
> I had some problem with some timeouts. My nfssserver init script runs
> more than 5 seconds. So the resource failed in version 2 style.
This might be the issues. I am going bootchart and see if this is causing
the problem.
>
> The most cases works fine with that configuration, failover etc works
> great. We want no autofailback, so it's off.
>
I am required, no need to ask why it's just how it is, to have failback
working. One of the reasons is we are running this in a non-homogenous
cluster. It is strictly for disaster situations. I would have it different
but that is not my call. Thank you for your reply. I think you may have
helped greatly. I will let you know!
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
