Re: [Pacemaker] Putting a node to standby stops resources on the other node
Hi' I got that explanation offline of the mailing list as well :-) :-). Still thanks a very lot for the explanation. Best regards. Thomas. Fra: Jake Smith Til: Thomas Lorenzen ; The Pacemaker cluster resource manager Sendt: 21:04 onsdag den 20. februar 2013 Emne: Re: [Pacemaker] Putting a node to standby stops resources on the other node Your rule: colocation backup_not_with_master -inf: master backup requires that the cluster knows where backup is running to be able to place master. Since backup wasn't running it couldn't put master anywhere. Now: colocation backup_not_with_master -inf: backup master Master can run anywhere and backup must only run where master isn't. ___ 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
Re: [Pacemaker] Putting a node to standby stops resources on the other node
- Original Message - > From: "Thomas Lorenzen" > To: m...@sys4.de, "The Pacemaker cluster resource manager" > > Sent: Wednesday, February 20, 2013 2:39:15 PM > Subject: Re: [Pacemaker] Putting a node to standby stops resources on > the other node > Hi' > Bingo, that seems to work. Still, I find it somewhat unexpected, that > the order matters in such a colocation rule. Anyway, things now > seems to work as expected. Thanks a very lot. Your rule: colocation backup_not_with_master -inf: master backup requires that the cluster knows where backup is running to be able to place master. Since backup wasn't running it couldn't put master anywhere. Now: colocation backup_not_with_master -inf: backup master Master can run anywhere and backup must only run where master isn't. HTH Jake > Best regards. > Thomas. > Fra: Michael Schwartzkopff > Til: Thomas Lorenzen ; The Pacemaker cluster > resource manager > Sendt: 20:28 onsdag den 20. februar 2013 > Emne: Re: [Pacemaker] Putting a node to standby stops resources on > the other node > better: > colocation backup_not_with_master -inf: backup master > ___ > 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
Re: [Pacemaker] Putting a node to standby stops resources on the other node
Hi' Bingo, that seems to work. Still, I find it somewhat unexpected, that the order matters in such a colocation rule. Anyway, things now seems to work as expected. Thanks a very lot. Best regards. Thomas. Fra: Michael Schwartzkopff Til: Thomas Lorenzen ; The Pacemaker cluster resource manager Sendt: 20:28 onsdag den 20. februar 2013 Emne: Re: [Pacemaker] Putting a node to standby stops resources on the other node better: colocation backup_not_with_master -inf: backup master___ 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
Re: [Pacemaker] Putting a node to standby stops resources on the other node
Am Mittwoch, 20. Februar 2013, 17:54:49 schrieb Thomas Lorenzen: > Hi' > > My setup is a simple two cluster solution intended for a "master" "backup" > framework, where I would like to assign floating ip addresses on two > network interfaces for both the master and the backup. > > The most cricial part of the config is as follows. > > group backup backup_ip_intern backup_ip_extern > > group master master_ip_intern master_ip_extern > colocation backupgroup_cannot_live_together_with_mastergroup -inf: master better: colocation backup_not_with_master -inf: backup master > backup order backupgroup_start_after_mastergroup inf: master backup > > > So, two groups are defined, each containing two primitives, which > essentially just assigns floating ip addresses to two different network > interfaces. The colocation rule is intended to make sure, that the groups > does not end up on the same node. The order rule is intended to make sure, > that the master group and not the backup group gets started, if only a > single node is available. Depeding on the verison of pacemaker you could also do this with utilizations and priorities. So you would not need the order and colocation constraint at all. Greetings. -- Mit freundlichen Grüßen, Michael Schwartzkopff -- [*] sys4 AG http://sys4.de, +49 (89) 30 90 46 64 Franziskanerstraße 15, 81669 München Sitz der Gesellschaft: München, Amtsgericht München: HRB 199263 Vorstand: Patrick Ben Koetter, Axel von der Ohe, Marc Schiffbauer Aufsichtsratsvorsitzender: Joerg Heidrich___ 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
Re: [Pacemaker] Putting a node to standby stops resources on the other node
Hi' Sorry, I do have that setting active. Properties are as follows. property $id="cib-bootstrap-options" \ dc-version="1.1.6-9971ebba4494012a93c03b40a2c58ec0eb60f50c" \ cluster-infrastructure="Heartbeat" \ expected-quorum-votes="2" \ stonith-enabled="false" \ no-quorum-policy="ignore" Best regards. Thomas. Fra: Michael Schwartzkopff Til: Thomas Lorenzen ; The Pacemaker cluster resource manager Sendt: 20:11 onsdag den 20. februar 2013 Emne: Re: [Pacemaker] Putting a node to standby stops resources on the other node Perhaps you do NOT have property no-quorum-policy=ignore___ 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
Re: [Pacemaker] Putting a node to standby stops resources on the other node
Am Mittwoch, 20. Februar 2013, 17:54:49 schrieb Thomas Lorenzen: > Hi' > > My setup is a simple two cluster solution intended for a "master" "backup" > framework, where I would like to assign floating ip addresses on two > network interfaces for both the master and the backup. > > The most cricial part of the config is as follows. > > group backup backup_ip_intern backup_ip_extern > > group master master_ip_intern master_ip_extern > colocation backupgroup_cannot_live_together_with_mastergroup -inf: master > backup order backupgroup_start_after_mastergroup inf: master backup > > > So, two groups are defined, each containing two primitives, which > essentially just assigns floating ip addresses to two different network > interfaces. The colocation rule is intended to make sure, that the groups > does not end up on the same node. The order rule is intended to make sure, > that the master group and not the backup group gets started, if only a > single node is available. > > The configuration works all fine with both nodes active. Now, when I put a > single node to standby, what happens is, that the resources on both nodes > are stopped, which is obviously not the intention. The expectation is of > course, that (1) the master group will run, when only a single node is > available, and (2) both master group and backup group will run on separate > nodes, when both nodes are available. > > Most probably the solution is trivial, but I simply can not get my mind > right on this. Any help will surely be appreciated. > > Best regards. > > Thomas. Perhaps you do NOT have property no-quorum-policy=ignore -- Mit freundlichen Grüßen, Michael Schwartzkopff -- [*] sys4 AG http://sys4.de, +49 (89) 30 90 46 64 Franziskanerstraße 15, 81669 München Sitz der Gesellschaft: München, Amtsgericht München: HRB 199263 Vorstand: Patrick Ben Koetter, Axel von der Ohe, Marc Schiffbauer Aufsichtsratsvorsitzender: Joerg Heidrich___ 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
