[Linux-HA] Centos 4.4 heartbeat 2.0.7-1.c4 and IPaddr2 problem

[Andrew Beekhof]

On 11/27/06, Toni <list-toni at ntg.fi> wrote:
> Hi,
>
> I have a problem after upgrading to heartbeat 2.0.7 with Streamline High
> Availability and Load Balancing setup:
>
>    http://www.ultramonkey.org/3/topologies/sl-ha-lb-overview.html
>
> The problem is not between two nodes, but with the public IP address I
> have configured to lo interface because it is never moved from lo to
> eth0 after starting heartbeat. This worked fine with v1.
>
> I have shutdown node2.domain.tld node to be sure it is not running
> heartbeat and reserving IP address.
>
> After weekend of investigating and trying different configs I'm pretty
> clueless if this is a problem with IPaddr2 script or just with my configs.
>
> CentOS 4.4  Linux 2.6.9-42.0.3.ELsmp
>
> Using RPM packages from centos extras:
>
>    heartbeat-stonith-2.0.7-1.c4
>    heartbeat-pils-2.0.7-1.c4
>    heartbeat-2.0.7-1.c4
>    heartbeat-ldirectord-2.0.7-1.c4
>
> Previously I used the old v1 centos extras RPM packages:
>
>    heartbeat.i386 1.2.3.cvs.20050927-1.centos4
>    heartbeat-ldirectord.i386 1.2.3.cvs.20050927-1.centos4
>    heartbeat-stonith.i386 1.2.3.cvs.20050927-1.centos4
>    heartbeat-pils.i386 1.2.3.cvs.20050927-1.centos4
>
> I've tried both v1 style haresources config and cib.xml. Same problem,
> though I use haresources2cib.py to create my cib.xml. Running crm_verify
> does not report errors.
>
> ha.cf:
> --- clip ---
> use_logd no
> crm no
> debug 1
> debugfile /var/log/ha-debug
> bcast eth1
> keepalive 10
> warntime 20
> deadtime 30
> initdead 60
> udpport 694
> auto_failback on
> node node1.domain.tld
> node node2.domain.tld
> respawn hacluster /usr/lib/heartbeat/ipfail
> apiauth ipfail gid=haclient uid=hacluster
> --- clip ---
>
> haresources (try #1):
> --- clip ---
> node1.domain.tld \
>         LVSSyncDaemonSwap::master \
>         ldirectord::ldirectord.cf \
>         IPaddr2::194.100.249.28/24/eth0/194.100.249.255

from IPaddr2

  echo "usage: $0 ip-address[/netmaskbits[/interface[:label][/broadcast]]]"


you might also want to try the CIDR netmask form (24) as there have
been bugs dealing with the dot-notation and I'm not 100% sure when
they were introduced

> --- clip ---
>
> haresources (try #2):
> --- clip ---
> node1.domain.tld 123.123.123.26 ldirectord::ldirectord.cf
> --- clip ---
>
> Could somebody explain if LVSSyncDaemonSwap::master is needed or not
> with v2 and this kind of setup? I've also tried with IPaddr, but I think
> it can't be used with this kind of setup (is it intended to move IP
> addresses from interface to another?)
>
> This is the IPaddr2 error I keep getting every time:
> --- clip ---
> ip-request-resp[5755]:  2006/11/25_22:03:49 received ip-request-resp
> LVSSyncDaemonSwap::master OK yes
> ResourceManager[5770]:  2006/11/25_22:03:49 info: Acquiring resource
> group: radon.nettitieto.fi LVSSyncDaemonSwap::master
> ldirectord::ldirectord.cf IPaddr2::123.123.123.26
> ResourceManager[5770]:  2006/11/25_22:03:49 info: Running
> /etc/ha.d/resource.d/LVSSyncDaemonSwap master start
> ResourceManager[5770]:  2006/11/25_22:03:49 debug: Starting
> /etc/ha.d/resource.d/LVSSyncDaemonSwap master start
> LVSSyncDaemonSwap[5832]:        2006/11/25_22:03:49 info:
> ipvs_syncbackup down
> LVSSyncDaemonSwap[5832]:        2006/11/25_22:03:49 info: ipvs_syncmaster up
> LVSSyncDaemonSwap[5832]:        2006/11/25_22:03:49 info:
> ipvs_syncmaster obtained
> ResourceManager[5770]:  2006/11/25_22:03:49 debug:
> /etc/ha.d/resource.d/LVSSyncDaemonSwap master start done. RC=0
> ldirectord is stopped for /etc/ha.d/ldirectord.cf
> ResourceManager[5770]:  2006/11/25_22:03:49 info: Running
> /etc/ha.d/resource.d/ldirectord ldirectord.cf start
> ResourceManager[5770]:  2006/11/25_22:03:49 debug: Starting
> /etc/ha.d/resource.d/ldirectord ldirectord.cf start
> ResourceManager[5770]:  2006/11/25_22:03:50 debug:
> /etc/ha.d/resource.d/ldirectord ldirectord.cf start done. RC=0
> /usr/lib/ocf/resource.d/heartbeat/IPaddr2: line 552: return: partial:
> numeric argument required
> /usr/lib/ocf/resource.d/heartbeat/IPaddr2: line 815: [: =: unary
> operator expected
> IPaddr2[5929]:  2006/11/25_22:03:50 INFO: IPaddr2 Resource is stopped
> ResourceManager[5770]:  2006/11/25_22:03:50 info: Running
> /etc/ha.d/resource.d/IPaddr2 123.123.123.26 start
> ResourceManager[5770]:  2006/11/25_22:03:50 debug: Starting
> /etc/ha.d/resource.d/IPaddr2 123.123.123.26 start
> /usr/lib/ocf/resource.d/heartbeat/IPaddr2: line 552: return: partial:
> numeric argument required
> IPaddr2[6157]:  2006/11/25_22:03:50 INFO: /usr/lib/heartbeat/send_arp -i
> 200 -r 5 -p /var/run/heartbeat/rsctmp/send_arp/send_arp-123.123.123.26
> eth0 123.123.123.26 auto 123.123.123.26 ffffffffffff
> IPaddr2[6070]:  2006/11/25_22:03:50 INFO: IPaddr2 Success
> INFO: IPaddr2 Success
> ResourceManager[5770]:  2006/11/25_22:03:50 debug:
> /etc/ha.d/resource.d/IPaddr2 123.123.123.26 start done. RC=0
> heartbeat[5648]: 2006/11/25_22:03:50 info: Exiting ip-request-resp
> process 5755 returned rc 0.
> heartbeat[5648]: 2006/11/25_22:03:50 debug: RscMgmtProc
> 'ip-request-resp' exited code 0
> heartbeat[5648]: 2006/11/25_22:03:50 info: AnnounceTakeover(local 1,
> foreign 1, reason 'ip-request-resp' (1))
> heartbeat[5648]: 2006/11/25_22:03:50 WARN: G_SIG_dispatch: Dispatch
> function for SIGCHLD took too long to execute: 70 ms (> 10 ms) (GSource:
> 0x9480428)
> heartbeat[5648]: 2006/11/25_22:03:58 info: Local Resource acquisition
> completed. (none)
> heartbeat[5648]: 2006/11/25_22:03:58 info: local resource transition
> completed.
> --- clip ---
>
> I didn't include ldirectord.cf at this point because I think it is not
> the problem here?
>
>
> Best regards,
>
>    ++Toni
> _______________________________________________
> Linux-HA mailing list
> Linux-HA at lists.linux-ha.org
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
>