Does anyone know of a User Group for Sun Cluster?
in [Solaris]
|
From: ITguy on 27 Jul 2010 23:10 > The one problem we've have is with patching. The only document we've > been able to find that documents the patching process for Sun Cluster > with HA-Zones does everything in single user mode with SC disabled. > We have a small 2 node Test cluster (test for our use, not user > testing), that we've been playing with to develop a procedure to patch > SC with HA-Zones using LiveUpgrade. We're almost have all the steps > worked out, but we've had very little luck working with Oracle/Sun to > get them to say of this will be a supported procedure or not. Patching clustered zones does suck. I don't know of a way to do it without significant down time for the zones. > If you are using Sun Cluster with HA-Zones, what procedures do you > follow to patch your nodes/zones? Would you mind sharing your general procedure? My thought would be something like: 1) Fail-over HA zone(s) to node "a" 2) Patch node "b" and rejoin cluster 3) Force HA zone(s) into detached state on server "b" by changing "installed" to "configured" in /etc/zones/index 4) Disable HA zone resource(s) - service outage begins 5) Fail HA resource groups to node "b" - HA zone(s) still disabled 6) Attach HA zones on node "b" with "-u" option to patch zones 7) Enable HA zone resource(s) - end service outage (hopefully!) 8) Patch node "a" and rejoin cluster Thoughts?
From: FatAnt on 28 Jul 2010 07:25 Tom wrote: > If you are using Sun Cluster with HA-Zones, what procedures do you > follow to patch your nodes/zones? I do not recommend to use liveupgrade. I'm agree with ITguy... Note for the node must be patched (node A): - disable all resource and devices group for the disks attached to zones - clnode evacuate A - node A: boot -sx - all zones on node A are disabled (the patching procedure that deals with turn on the zones) - the disk group (for zones) must be imported on the node A (es. with veritas VM) - patching... - reboot A. now A is in cluster - IMPORTANT: detach any ha-zone on node B - clnode evacuate B - node B: boot -sx - patching... NOTE: and of course detach one of two/three mirror disks in case you need to use the patching rollback I suggest you to see in the document "Sun Cluster Upgrade Guide for Solaris OS" the procedure "Performing a Rolling Upgrade " http://docs.sun.com/app/docs/doc/820-7357 PS: I'm sorry for my terrible english, I hope to have been sufficiently understood :-)
From: Tom on 29 Jul 2010 15:56 On Jul 27, 11:10 pm, ITguy <southa...(a)gmail.com> wrote: > > The one problem we've have is with patching. The only document we've > > been able to find that documents the patching process for Sun Cluster > > with HA-Zones does everything in single user mode with SC disabled. > > We have a small 2 node Test cluster (test for our use, not user > > testing), that we've been playing with to develop a procedure to patch > > SC with HA-Zones using LiveUpgrade. We're almost have all the steps > > worked out, but we've had very little luck working with Oracle/Sun to > > get them to say of this will be a supported procedure or not. > > Patching clustered zones does suck. I don't know of a way to do it > without significant down time for the zones. > > > If you are using Sun Cluster with HA-Zones, what procedures do you > > follow to patch your nodes/zones? > > Would you mind sharing your general procedure? My thought would be > something like: > 1) Fail-over HA zone(s) to node "a" > 2) Patch node "b" and rejoin cluster > 3) Force HA zone(s) into detached state on server "b" by changing > "installed" to "configured" in /etc/zones/index > 4) Disable HA zone resource(s) - service outage begins > 5) Fail HA resource groups to node "b" - HA zone(s) still disabled > 6) Attach HA zones on node "b" with "-u" option to patch zones > 7) Enable HA zone resource(s) - end service outage (hopefully!) > 8) Patch node "a" and rejoin cluster > > Thoughts? Take a look at "http://docs.sun.com/app/docs/doc/819-2971/6n57mi2g0". It works well, except for all of the downtime when patching the HA- Zones.
From: Tom on 29 Jul 2010 16:14 I have no problem with using LiveUpgrade (This would prevent significant downtime for patching the HA-Zones, and gives you a back out option that you otherwise don't have). The problem I do have is the lack of information from Oracle/Sun on either: 1) the Best Practice procedure for using LiveUpgrade with HA-Zones. If LiveUpgrade is a viable option to patching Sun Cluster with HA- Zones, it would be nice to be told the correct steps to use in order to be fully supported. or 2) a warning to not use LiveUpgrade with HA-Zones. If LiveUpgrade is unsafe, and shouldn't be used, it would be nice to know that as well. By the way, your English is very good. On Jul 28, 7:25 am, FatAnt <fat...(a)blackhole.null> wrote: > Tom wrote: > > If you are using Sun Cluster with HA-Zones, what procedures do you > > follow to patch your nodes/zones? > > I do not recommend to use liveupgrade. Why?? This prevents significant downtime for patching the HA-Zones, and gives you backout options that you otherwise don't have? I have no problem with using LiveUpgrade. It isn't LiveUpgrade I have a problem with, but the lack of information from Oracle/Sun on the Best Practice procedure for using LiveUpgrade with HA-Zones, or a warning to not use LiveUpgrade with HA-Zones. > > I'm agree with ITguy... > > Note for the node must be patched (node A): > > - disable all resource and devices group for the disks attached to zones > - clnode evacuate A > - node A: boot -sx > - all zones on node A are disabled (the patching procedure that deals with > turn on the zones) > - the disk group (for zones) must be imported on the node A (es. with > veritas VM) > - patching... > > - reboot A. now A is in cluster > > - IMPORTANT: detach any ha-zone on node B > - clnode evacuate B > - node B: boot -sx > - patching... > > NOTE: and of course detach one of two/three mirror disks in case you need > to use the patching rollback > > I suggest you to see in the document "Sun Cluster Upgrade Guide for Solaris > OS" the procedure "Performing a Rolling Upgrade " > > http://docs.sun.com/app/docs/doc/820-7357 > > PS: I'm sorry for my terrible english, I hope to have been sufficiently > understood :-)
From: Tom on 29 Jul 2010 16:17 I have no problem with using LiveUpgrade (This would prevent significant downtime for patching the HA-Zones, and gives you a back out option that you otherwise don't have). The problem I do have is the lack of information from Oracle/Sun on either: 1) the Best Practice procedure for using LiveUpgrade with HA-Zones. If LiveUpgrade is a viable option to patching Sun Cluster with HA- Zones, it would be nice to be told the correct steps to use in order to be fully supported. or 2) a warning to not use LiveUpgrade with HA-Zones. If LiveUpgrade is unsafe, and shouldn't be used, it would be nice to know that as well. By the way, your English is very good. On Jul 28, 7:25 am, FatAnt <fat...(a)blackhole.null> wrote: > Tom wrote: > > If you are using Sun Cluster with HA-Zones, what procedures do you > > follow to patch your nodes/zones? > > I do not recommend to use liveupgrade. > > I'm agree with ITguy... > > Note for the node must be patched (node A): > > - disable all resource and devices group for the disks attached to zones > - clnode evacuate A > - node A: boot -sx > - all zones on node A are disabled (the patching procedure that deals with > turn on the zones) > - the disk group (for zones) must be imported on the node A (es. with > veritas VM) > - patching... > > - reboot A. now A is in cluster > > - IMPORTANT: detach any ha-zone on node B > - clnode evacuate B > - node B: boot -sx > - patching... > > NOTE: and of course detach one of two/three mirror disks in case you need > to use the patching rollback > > I suggest you to see in the document "Sun Cluster Upgrade Guide for Solaris > OS" the procedure "Performing a Rolling Upgrade " > > http://docs.sun.com/app/docs/doc/820-7357 > > PS: I'm sorry for my terrible english, I hope to have been sufficiently > understood :-)
First
|
Prev
|
Next
|
Last
Pages: 1 2 3 Prev: About Apache2 with RBAC Next: How can I find out if CPU throttling is enabled? |