the cluster configuration is a 2 node setup to run ha-nfs in a non-shared storage environment - except for the shared avs cluster database which is stored on an iscsi lun shared from my quorum server.
Using the work from Marcelo Leal - http://www.eall.com.br/blog/?p=75 i am using his agent to manage the avs failover between nodes Everything works fine upon initial setup however when i reboot node (a), the failover happens perfectly but when node (a) comes back, it rejoins the cluster and then appears to have no idea about the sndr >From your response it sounds like my use of AVS in a cluster is the problem which im rather surprised of the limitation?? I would of thought AVS would of been the perfect fit if i wanted to make a poor mans replicating SAN with 2 servers attached to some J4200 and if im providing NFS to my network that i could use clustering... Perhaps someone can explain the reasoning why i cant have 2 nodes in the same cluster replicating their disks via avs and why it works on initial setup?? Should there be data written to the shared avs cluster database which contains the same information as dscfg_local?? I have tried the iscsi method (http://hub.opensolaris.org/bin/download/Project+colorado/files/Whitepap er-OpenHAClusterOnOpenSolaris-external.pdf) which also works nicely however when a node fails its noticeable to the clients connected to the NFS that something has happend until zfs realises that one of the iscsi luns has gone missing... Here is a snippet of my ds.log Mar 18 15:27:39 ii: iiboot resume cluster tag <none> Mar 18 15:28:44 sndr: sndradm -E nas1 /dev/rdsk/c1t1d0s0 /dev/rdsk/c1t1d0s1 nas2 /dev/rdsk/c1t1d0s0 /dev/rdsk/c1t1d0s1 Successful Mar 18 15:28:44 sv: enabled /dev/rdsk/c1t1d0s0 Mar 18 15:28:44 sv: enabled /dev/rdsk/c1t1d0s1 Mar 18 15:29:16 sv: svboot: suspend /dev/rdsk/c1t1d0s0 Mar 18 15:29:16 sv: svboot: suspend /dev/rdsk/c1t1d0s1 Mar 18 15:29:16 scm: scmadm cache disable succeeded Mar 18 15:29:16 cfglockd: pid 5009 terminate on signal 15 Mar 18 15:31:20 cfglockd: new lock daemon, pid 593 Mar 18 15:31:20 cfglockd: pid 593 unexpected signal 18, ignoring BTW is there any active work being done on AVS for the upcoming release of opensolaris? Cheers for the help guys -----Original Message----- From: thomas.atk...@sun.com [mailto:thomas.atk...@sun.com] Sent: Thursday, 18 March 2010 10:55 PM To: Chris Anders Cc: storage-discuss@opensolaris.org Subject: Re: [storage-discuss] problems with AVS with cluster 3.2 Hi Chris, Check /var/adm/ds.log for error messages. Can you describe the cluster configuration? In your example, avs1 and avs2 cannot be part of the same cluster. avs1 can be a logical host in one cluster, but avs2 needs to be a logical host in a different cluster or a standalone machine. I would need more info but it sounds like when the node reboots, the logical host fails over but the storage is not failing over. Regards, Tom On 03/18/10 03:03, Chris Anders wrote: > Hi, im running latest solaris with AVS 4.0 and cluster 3.2 U3 and i > appear to have a problem with the cluster configuration shared databse > > > > I can setup my cluster fine, i can setup the replication fine however > when a node reboots the replication doesn't start again. > > > > When i run dsstat the return code is 0 but returns no output > > When i try to recreate the replication im told: > > > > Remote Mirror: avs1 /dev/rdsk/c8t1d0s0 /dev/rdsk/c8t1d0s1 avs2 > /dev/rdsk/c8t1d0s0 /dev/rdsk/c8t1d0s1 > sndradm: warning: SNDR: /dev/rdsk/c8t1d0s0 ==> /dev/rdsk/c8t1d0s0 not > already enabled > > > > If i run dscfg -l | grep -v "#" i see: > > scm: 128 64 - - - - - - - > > sndr: nas1 /dev/rdsk/c1t1d0s0 /dev/rdsk/c1t1d0s1 nas2 /dev/rdsk/c1t1d0s0 > /dev/rd > sk/c1t1d0s1 ip sync tank1 - setid=1; - > > sv: /dev/rdsk/c1t1d0s0 - - > > sv: /dev/rdsk/c1t1d0s1 - - > > dsvol: /dev/rdsk/c1t1d0s0 - sndr > > dsvol: /dev/rdsk/c1t1d0s1 - sndr > > > > > > yet if i run dscfg -l -s /dev/did/rdsk/d5s0 | grep -v "#" i see: > > setid: 1 setid-ctag > > > > ive tried this same setup using opensolaris only to hit the same wall.. > > > > am i missing something with AVS when using it in a cluster?? Ive only > ever used it in non-cluster mode where i have never had a problem... > > > > cheers for any help > > > > Chris Anders > > > ------------------------------------------------------------------------ > > _______________________________________________ > storage-discuss mailing list > storage-discuss@opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/storage-discuss _______________________________________________ storage-discuss mailing list storage-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/storage-discuss
