On Wed, Dec 21, 2011 at 6:01 PM, Nirmala S <nirma...@huawei.com> wrote: > Hi, > > > > This is a followup on earlier thread > (http://www.gossamer-threads.com/lists/linuxha/pacemaker/76705). > > > > My situation is somewhat similar. I need to a cluster which contains 3 kinds > of nodes – master, preferred slave, slave. Preferred slave is an entity that > becomes the master in case of switchover/failover. Master is the master for > pref_slave and pref_slave is master for other slaves. The master election is > easy – it is done by crm, all I need to do is use crm_master. > > > > But for the preferred slave, there needs to an election amongst existing > slaves. As of now I am using a variable in CIB with > pref_slave|pref_slave_score|temp_score. If temp_score is 0, then the slave > will update pref_slave and pref_slave_score and temp_score. If temp_score is > non-zero, then the node compares its score with pref_slave_score and updates > only if it is bigger. >
What is the difference between a preferred slave and a normal one? Sounds like what you really want is tri-state resources (instead of the existing bi-state) > > Now I have 2 problems > > Everytime I change the CIB(which I am doing in pre-promote), the event > (pre-promote) is getting retriggered. Is doing it during post-promote an option? > The event(pre-promote) is sent in parallel to all the slaves. So each slave > thinks temp_score is 0, and overwrites with its score. Is there any way to > serialize this using some sort of lock ? Basically, no. > Or is there a provision to store > cluster-wide attributes apart from CIB ? Also no. > Regards > > Nirmala > > > > > > This e-mail and attachments contain confidential information from HUAWEI, > which is intended only for the person or entity whose address is listed > above. Any use of the information contained herein in any way (including, > but not limited to, total or partial disclosure, reproduction, or > > dissemination) by persons other than the intended recipient's) is > prohibited. If you receive this e-mail in error, please notify the sender by > phone or email immediately and delete it! > > > > > _______________________________________________ > Pacemaker mailing list: Pacemaker@oss.clusterlabs.org > http://oss.clusterlabs.org/mailman/listinfo/pacemaker > > Project Home: http://www.clusterlabs.org > Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf > Bugs: http://bugs.clusterlabs.org > _______________________________________________ Pacemaker mailing list: Pacemaker@oss.clusterlabs.org http://oss.clusterlabs.org/mailman/listinfo/pacemaker Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org
