Am Donnerstag, den 11.12.2008, 15:27 +0100 schrieb Maros TIMKO:
> Thanks Thomas for very prompt answer.
> >> we are building Xen virtual machine failover system using heartbeat 2.1.3
> >> (v1 config style - 2 node cluster setup) on CentOS 5.2.
> >
> >Make your life easier and use hb2 with crm.
>
> Our problem is if we would later on want to use live migration, CRM does not
> support dual primary setup for DRBD.
ah?
# crm_mon -i 1
Refresh in 1s...
============
Last updated: Thu Dec 11 15:42:52 2008
Current DC: kzvxen2 (aa7f65f8-3aa6-4db4-a262-f52e798038d9)
2 Nodes configured.
19 Resources configured.
============
Node: kzvxen2 (aa7f65f8-3aa6-4db4-a262-f52e798038d9): online
Node: kzvxen1 (ad0cd857-d765-44f2-a17d-0b689b63774b): online
Clone Set: stonithclone
stonithclone:0 (stonith:external/ssh): Started kzvxen2
stonithclone:1 (stonith:external/ssh): Started kzvxen1
....cut.....
univ (heartbeat::ocf:Xen): Started kzvxen1
openqrm (heartbeat::ocf:Xen): Started kzvxen2
ldap1 (heartbeat::ocf:Xen): Started kzvxen1
ldap2 (heartbeat::ocf:Xen): Started kzvxen2
lb (heartbeat::ocf:Xen): Started kzvxen2
...... cut.......
# cat /proc/drbd
version: 8.0.13 (api:86/proto:86)
GIT-hash: ee3ad77563d2e87171a3da17cc002ddfd1677dbe build by
[EMAIL PROTECTED], 2008-08-04 15:28:07
0: cs:Connected st:Primary/Primary ds:UpToDate/UpToDate C r---
ns:0 nr:25425620 dw:25425620 dr:44 al:0 bm:204 lo:0 pe:0 ua:0 ap:0
resync: used:0/61 hits:60980 misses:204 starving:0 dirty:0 changed:204
act_log: used:0/257 hits:0 misses:0 starving:0 dirty:0 changed:0
1: cs:Connected st:Primary/Primary ds:UpToDate/UpToDate C r---
ns:1469916 nr:86016 dw:1555932 dr:91701 al:91 bm:18 lo:0 pe:0 ua:0
ap:0
resync: used:0/61 hits:5358 misses:18 starving:0 dirty:0 changed:18
act_log: used:0/257 hits:367388 misses:91 starving:0 dirty:0 changed:91
> >- Install 2 domus for openldap, setup ldap-replication
> >- use the xen-domu as cluster-ressources with constraints, so you can
> >say ldap1 rans preffered on XENBOX1 and ldap on xenbox2
> >
>
> Here the problem is we are using LDAP for Dom0 user authenication too.
> Thus without running domain we would not have LDAP and we would not
> be able to log in. That's why LDAP must be running on both nodes all the time
> not only as an HA resource.
I meant you _have to_ run 2 ressource: ldap1 and ldap2. Work with
constraints and place e.g. ldap1 on Server1 and ldap2 on Server2
Fire up a load-balancer-ressource on dom0, which does round-robin for
ldap-requests.
I also put all these servers into xen-domus which makes handling easier
for me. BUt this does not work for you, since you have to have the
load-balancer online all-the-time.
>
> >if you prefer v1 this should help:
> >http://www.linuxjournal.com/article/5505
>
> It's a bit complicated in the sense of restarting services, but very good
> starting point. Thanks for pointing to this.
>
> >> Next solution would be to setup LDAP into a cluster. However, as far as
> >> I checked the docu, we need to specify IP address of slave and I am not
> >> sure how it works when we need to switchover.
> >
> >openLDAP 2.3 supports only
> >- Master/Slave-Replication
> >so you need heartbeat for HA
> >
> >if running 2.4 you can also the built-in mirror-mode (for active
> >server-to-server replication used in high-availability environments)
> >
>
> Well, CentOS distribution includes 2.3 version, we must keep it.
So - you have the built-in-replication which keeps your ldap synchronous
on all Servers. Put a load-balancer in front of the ldap-server and
you're done.
>
> >>
> >> Is there anyone experienced with such a problem?
> >
> >nope - running fine here with heartbeat and simple
> >master/slave-replication with service-IP
> >
> >
> Thanks.
Thomas
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
