Re: [suse-oracle] cannot mount shared file system on more than 3 nodes

Hi Diedrich,

did you have check if the firewall is blocking something?
Are all cluster.conf identical (md5sum /etc/ocfs2/cluster.conf)

cu
 Peter

 Fri, Nov 28, 2008 at 08:45:19AM +0100, Diedrich Ehlerding wrote:
> I am trying to set up a RAC cluster with 4 nodes. Due to SAP's 
> restriction, I cannot use ASM but must use ocfs2 filesystems.
> 
> The servers are setup with SLES10 SP2, kerel update -33. I get a 
> filesystem mounted on three nodes, but the fourth one display an error 
> messae.
> 
> The cluster is working: 
> DB02:~ # /etc/init.d/o2cb status
> Module "configfs": Loaded
> Filesystem "configfs": Mounted
> Module "ocfs2_nodemanager": Loaded
> Module "ocfs2_dlm": Loaded
> Module "ocfs2_dlmfs": Loaded
> Filesystem "ocfs2_dlmfs": Mounted
> Checking O2CB cluster ocfs2: Online
> Heartbeat dead threshold = 31
>   Network idle timeout: 30000
>   Network keepalive delay: 2000
>   Network reconnect delay: 2000
> Checking O2CB heartbeat: Active
> 
> (same on all 4 nodes).
> 
> I try to mount a ahred filesystem:
> 
> 
> DB02:~ # mount /dev/mapper/*ora* /mnt
> DB02:~ # mount | grep mnt
> /dev/mapper/SL1_1_oracle_0_1 on /mnt type ocfs2 
> (rw,_netdev,heartbeat=local)
> DB02:~ # ssh DB03
> Last login: Fri Nov 28 00:26:30 2008 from 10.0.1.40
> DB03:~ # mount | grep mnt
> DB03:~ # mount /dev/mapper/*ora* /mnt
> DB03:~ # mount | grep mnt
> /dev/mapper/SL1_1_oracle_0_1 on /mnt type ocfs2 
> (rw,_netdev,heartbeat=local)
> DB03:~ # ssh DB04
> Last login: Thu Nov 27 23:33:54 2008 from 10.0.1.40
> DB04:~ # mount /dev/mapper/*ora* /mnt
> DB04:~ # mount | grep mnt
> /dev/mapper/SL1_1_oracle_0_1 on /mnt type ocfs2 
> (rw,_netdev,heartbeat=local)
> DB04:~ # ssh DB05
> Last login: Thu Nov 27 23:32:31 2008 from 10.0.1.40
> DB05:~ # mount /dev/mapper/*ora* /mnt
> mount.ocfs2: Transport endpoint is not connected while mounting 
> /dev/mapper/SL1_1_oracle_0_1 on /mnt. Check 'dmesg' for more 
> information on this error.
> 
> dmesg says:
> o2net: connected to node db02 (num 0) at 10.0.1.2:7777
> o2net: connected to node db03 (num 1) at 10.0.1.3:7777
> (10001,2):o2net_connect_expired:1637 ERROR: no connection established 
> with node 2 after 30.0 seconds, giving up and returning errors.
> (12500,2):dlm_request_join:1033 ERROR: status = -107
> (12500,2):dlm_try_to_join_domain:1208 ERROR: status = -107
> (12500,2):dlm_join_domain:1487 ERROR: status = -107
> (12500,2):dlm_register_domain:1734 ERROR: status = -107
> (12500,2):ocfs2_dlm_init:2662 ERROR: status = -107
> (12500,2):ocfs2_mount_volume:1255 ERROR: status = -107
> ocfs2: Unmounting device (253,19) on (node 3)
> o2net: no longer connected to node db02 (num 0) at 10.0.1.2:7777
> o2net: no longer connected to node db03 (num 1) at 10.0.1.3:7777
> 
> 
> The cluster was setup by ocfs2console, mainly default values:
> 
> DB05:~ # cat /etc/ocfs2/cluster.conf
> node:
>         ip_port = 7777
>         ip_address = 10.0.1.2
>         number = 0
>         name = db02
>         cluster = ocfs2
> 
> node:
>         ip_port = 7777
>         ip_address = 10.0.1.3
>         number = 1
>         name = db03
>         cluster = ocfs2
> 
> node:
>         ip_port = 7777
>         ip_address = 10.0.1.4
>         number = 2
>         name = db04
>         cluster = ocfs2
> 
> node:
>         ip_port = 7777
>         ip_address = 10.0.1.5
>         number = 3
>         name = db05
>         cluster = ocfs2
> 
> cluster:
>         node_count = 5
>         name = ocfs2
> 
> DB05:~ #
> 
> 
> as you se, I experimented with increasing  the numer of cluster noes 
> (originally 4), but that doenst change anything. 
> 
> The messages seem to indicate to me that there is sme ind of network 
> problem. However, all the nodes can see one another, and ssh cnnections 
> show no problem.
> 
> 
> 
> 
> 
> 
> 
> 
> -- 
> Diedrich Ehlerding, Fujitsu Siemens Computers GmbH, ISC SOP PS N/O
> Hildesheimer Str 25, D-30880 Laatzen
> Fon +49 511 8489-1806, Fax -251806, Mobil +49 173 2464758
> Firmenangaben: http://www.fujitsu-siemens.de/imprint.html
> _______________________________________________
> suse-oracle mailing list
> [email protected]
> http://listx.novell.com/mailman/listinfo/suse-oracle

-- 
Peter Schinagl - Technical ISV Relationship Manager
SUSE LINUX Products GmbH,Maxfeldstrasse 5, 90409 Nürnberg,Germany
Tel  : +49-911-74053-684           Mobil: +49-179-1308082
Email: [email protected]
-----------------------------------------------------------------
SUSE LINUX Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg)
-----------------------------------------------------------------
_______________________________________________
suse-oracle mailing list
[email protected]
http://listx.novell.com/mailman/listinfo/suse-oracle