[Glass] High Availability in GemStone

[Dale Henrichs]

Just a nit...

If you have to replay tranlogs to bring the stone up-to-date, that is more
of a warm standby than hot standby ... the hot standby is designed to keep
the standby stone within a few transactions of the master ...

Prior to the introduction of the hot standby functionality, warm standby
was the only option and was/is used by a number of commercial customers

Dale


On Thu, Feb 20, 2014 at 3:08 PM, James Foster <
james.foster at gemtalksystems.com> wrote:

> Bruno,
>
> I'm not that familiar with VMware fail-over but a quick Google search took
> me to
> http://pubs.vmware.com/vsphere-4-esx-vcenter/index.jsp?topic=/com.vmware.vsphere.availability.doc_41/c_useha_works.html
> .
>
> One way I'd approach this is to look at separating the redundant hosts
> (CPU/RAM) from the redundant disks. If you had a disk system (with RAID,
> etc.) that could be switched easily between hosts, then when a machine
> failed you could restart on another machine that would use *the exact
> same extents and transaction logs* as the original. This would avoid
> GemStone altogether when it comes to the "hot standby" issue. The only
> delay would be in replaying the transactions since the last checkpoint--and
> checkpoint frequency is configurable. See
> http://pubs.vmware.com/vsphere-51/index.jsp?topic=%2Fcom.vmware.vsphere.storage.doc%2FGUID-52DC7277-5321-4BB5-86B4-D73D258F6529.html<http://pubs.vmware.com/vsphere-51/index.jsp?topic=/com.vmware.vsphere.storage.doc/GUID-52DC7277-5321-4BB5-86B4-D73D258F6529.html> for
> a discussion of "Sharing a VMFS Datastore Across Hosts."
>
> Again, as you suggest, with this approach it is no longer a GemStone DBA
> problem but a datacenter administrator problem.
>
> James Foster
> Director of Operations
> GemTalk Systems, LLC
> +1 503 766 4714 (voice & fax)
> James.Foster at GemTalkSystems.com
>
> On Feb 20, 2014, at 1:59 PM, BrunoBB <smalltalk at adinet.com.uy> wrote:
>
> James,
>
> Let suppose that a company has a VM Ware virtual datacenter with lot of
> Linux running on it.
>
> Let say Node-1 is the master and Node-2 slave, and Node-1 is copied (at
> VMWare level, maybe at memory level) to Node-2.
>
> If Node-1 crash then Node-2 take the requests. If it is possible at VWare
> level maybe it needs some GemStone configuration or not ?
>
> (i do not know if it possible i will found out with a VM Ware
> administrator)
>
> If this is possible then we have "traslated the problem" to the datacenter
> administrator :)
>
> Regards,
> Bruno
>
>
>
> --
> View this message in context:
> http://forum.world.st/High-Availability-in-GemStone-tp4745211p4745382.html
> Sent from the GLASS mailing list archive at Nabble.com.
> _______________________________________________
> Glass mailing list
> Glass at lists.gemtalksystems.com
> http://lists.gemtalksystems.com/mailman/listinfo/glass
>
>
>
> _______________________________________________
> Glass mailing list
> Glass at lists.gemtalksystems.com
> http://lists.gemtalksystems.com/mailman/listinfo/glass
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gemtalksystems.com/mailman/private/glass/attachments/20140220/be7dd00b/attachment-0001.html>