[ZODB-Dev] Clustering ZEO Server with drbd
Hi folks, I'm looking at making a scalable zope infrastructure, I can scale the ZEO clients and mysql, the only issue is the ZEO server. I have used drbd and heartbeat in the past to create a HA lamp stack. From a technical standpoint, is there any reason I could not use drbd to cluster the ZEO server? Any thoughts would be appreciated. ___ For more information about ZODB, see the ZODB Wiki: http://www.zope.org/Wikis/ZODB/ ZODB-Dev mailing list - ZODB-Dev@zope.org http://mail.zope.org/mailman/listinfo/zodb-dev
Re: [ZODB-Dev] Clustering ZEO Server with drbd
If you setup drbd and heartbeat , the spare will take over the floating ip from the zeo server and restart the service. drbd is simply doing a block level replication of the zeo server file system where the zeo data is stored. The question is what files needs to be replicated so that when the the service is restarted on the spare, all the files needed by the zeo server to operate are there? I'm trying to ensure from a system architecture standpoint I'm not missing anything. Christian Theune [EMAIL PROTECTED] wrote: [EMAIL PROTECTED] wrote: Hi folks, I'm looking at making a scalable zope infrastructure, I can scale the ZEO clients and mysql, the only issue is the ZEO server. I have used drbd and heartbeat in the past to create a HA lamp stack. From a technical standpoint, is there any reason I could not use drbd to cluster the ZEO server? Any thoughts would be appreciated.As far as I understand, the difference to ZRS (which is a solution to scale ZEO servers, e.g. for HA) that you will need to restart the ZEO server when switching a hot-spare ZEO server to "live-mode", because it doesn't notice when the underlying storage file is changed by somebody else.Christian-- gocept gmbh co. kg - forsterstra�e 29 - 06112 halle/saale - germanywww.gocept.com - [EMAIL PROTECTED] - phone +49 345 122 9889 7 -fax +49 345 122 9889 1 - zope and plone consulting and development___ For more information about ZODB, see the ZODB Wiki: http://www.zope.org/Wikis/ZODB/ ZODB-Dev mailing list - ZODB-Dev@zope.org http://mail.zope.org/mailman/listinfo/zodb-dev
Re: [ZODB-Dev] Clustering ZEO Server with drbd
I am planning a ERP5 deployment with ZEO. I have a working HA configuration on Centos 4.3 (should work without changes on RHEL4 but I have not tested) and am willing to contribute it all to the community if I can get help clustering ZEO. I will write a howto and contribute back to the zope documentation. Anybody interested?Benji York [EMAIL PROTECTED] wrote: Christian Theune wrote: In the end, this might not be exactly HA because you'll have a period of time that isn't covered and may stretch for quite a while.An interesting idea: use drdb to create an HA pair for the read/write ZEO service, and also use a read-only ZRS secondary. That way when the primary r/w ZEO server went down, read transactions could still complete until the second r/w ZEO server could start.-- Benji YorkSenior Software EngineerZope Corporation___ For more information about ZODB, see the ZODB Wiki: http://www.zope.org/Wikis/ZODB/ ZODB-Dev mailing list - ZODB-Dev@zope.org http://mail.zope.org/mailman/listinfo/zodb-dev
Re: [ZODB-Dev] Clustering ZEO Server with drbd
I have two IBM xseries servers dedicated to this project at my datacenter. I am looking for anyone with zeo experience and expertise interested in working on this. I think this will make a HA zope infrastructure more attractive to the enterprise. I am also willing to do all the documentation and QA testing. Dieter Maurer [EMAIL PROTECTED] wrote: [EMAIL PROTECTED] wrote at 2006-5-28 19:30 -0700:Hi folks, I'm looking at making a scalable zope infrastructure, I can scale the ZEO clients and mysql, the only issue is the ZEO server. I have used drbd and heartbeat in the past to create a HA lamp stack. From a technical standpoint, is there any reason I could not use drbd to cluster the ZEO server? Any thoughts would be appreciated. I think it would be feasable. Of course, you may miss a few transactions...Case must be taken to handle packing correctly. "repozo" might helpin mirroring the storage.-- Dieter___ For more information about ZODB, see the ZODB Wiki: http://www.zope.org/Wikis/ZODB/ ZODB-Dev mailing list - ZODB-Dev@zope.org http://mail.zope.org/mailman/listinfo/zodb-dev
Re: [ZODB-Dev] Clustering ZEO Server with drbd
I would be very helpful if someone could summarize what needs to be done to get this to work. I don't have any experience with the ZEO server so any input would be greatly appreciated. Dieter Maurer [EMAIL PROTECTED] wrote: Christian Theune wrote at 2006-5-29 18:02 +0200: ...(Re)starting the ZEO server when it is needed might take a while on large databases if the index file happens to be broken.It would not be difficult to implement an incrementalindex determination. This would use a slghtly out of data index fileand then analyse the new (with respect to the index) entries inthe storage.-- Dieter___ For more information about ZODB, see the ZODB Wiki: http://www.zope.org/Wikis/ZODB/ ZODB-Dev mailing list - ZODB-Dev@zope.org http://mail.zope.org/mailman/listinfo/zodb-dev
