Re: [Linux-HA] How to start Pacemaker in unmanaged mode ?
On 5/11/2011 at 06:49 PM, Dominik Klein wrote: > On 05/11/2011 10:24 AM, Alain.Moulle wrote: > > Hi Dominik, > > I just have tried again : > > "service corosync stop" on both nodes node1 & node2 > > remove the cib.xml on node2 > > vi cib.xml on node1 > > set the property maintenance-mode=true > > ( > name="maintenance-mode" value="true"/>) > > wq! > > start Pacemaker on node1 and on node2 > > On node1 : both nodes remain UNCLEAN(offline) at vitam eternam > > On node2 : crm_mon : Attempting connection to the cluster. at vitam > > eternam > > Moreover , the added line for maintenance-mode=true has been removed by > > Pacemaker/corosync in the cib.xml > > > > I tried long time ago to do vi and change some values in the cib.xml, > > and each time > > I did that, Pacemaker never start correctly again and I had to > > reconfigure all the > > things from scratch so that it start again... > > I have to admit I only did this back in heartbeat times and so it seems > something changed regarding this. See: http://www.clusterlabs.org/doc/en-US/Pacemaker/1.1/html/Pacemaker_Explained/s-config-updates.html Specifically the bit that says: There are three basic rules for updating the cluster configuration: Rule 1 - Never edit the cib.xml file manually. Ever. I'm not making this up. Rule 2 - Read Rule 1 again. Rule 3 - The cluster will notice if you ignored rules 1 & 2 and refuse to use the configuration. > So ... Just had a look at a cluster of mine and there are backup copies > and .sig files for each cib.xml version. > > I don't know what pacemaker will do if you just > > a) removed the history and .sig file, so only cib.xml is in place > or > b) replaced the (apparently md5) checksum in .sig > > Worth a shot I think. Bearing in mind the above documentation, and a further admonishment that if you break it you get to keep both pieces :) you can generate one of those digests by running: # cibadmin --xml-file cib.xml --md5-sum Good luck, Tim -- Tim Serong Senior Clustering Engineer, OPS Engineering, Novell Inc. ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
Re: [Linux-HA] How to start Pacemaker in unmanaged mode ?
On 05/11/2011 10:24 AM, Alain.Moulle wrote: > Hi Dominik, > I just have tried again : > "service corosync stop" on both nodes node1 & node2 > remove the cib.xml on node2 > vi cib.xml on node1 > set the property maintenance-mode=true > ( name="maintenance-mode" value="true"/>) > wq! > start Pacemaker on node1 and on node2 > On node1 : both nodes remain UNCLEAN(offline) at vitam eternam > On node2 : crm_mon : Attempting connection to the cluster. at vitam > eternam > Moreover , the added line for maintenance-mode=true has been removed by > Pacemaker/corosync in the cib.xml > > I tried long time ago to do vi and change some values in the cib.xml, > and each time > I did that, Pacemaker never start correctly again and I had to > reconfigure all the > things from scratch so that it start again... I have to admit I only did this back in heartbeat times and so it seems something changed regarding this. So ... Just had a look at a cluster of mine and there are backup copies and .sig files for each cib.xml version. I don't know what pacemaker will do if you just a) removed the history and .sig file, so only cib.xml is in place or b) replaced the (apparently md5) checksum in .sig Worth a shot I think. Dominik ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
Re: [Linux-HA] How to start Pacemaker in unmanaged mode ?
Hi Dominik, I just have tried again : "service corosync stop" on both nodes node1 & node2 remove the cib.xml on node2 vi cib.xml on node1 set the property maintenance-mode=true () wq! start Pacemaker on node1 and on node2 On node1 : both nodes remain UNCLEAN(offline) at vitam eternam On node2 : crm_mon : Attempting connection to the cluster. at vitam eternam Moreover , the added line for maintenance-mode=true has been removed by Pacemaker/corosync in the cib.xml I tried long time ago to do vi and change some values in the cib.xml, and each time I did that, Pacemaker never start correctly again and I had to reconfigure all the things from scratch so that it start again... Alain Dominik Klein a écrit : > Correct me if I'm wrong but I strongly think the following would work: > > Since you want to _start_ pacemaker in unmanaged mode, I expect all your > nodes to be offline. Then delete cib.xml on all nodes but one. > > On the one remaining, edit cib.xml and put your configuration > is_managed=false there. > > Then start all nodes. Don't see why this shouldn't work. > > Regards > Dominik > > On 05/10/2011 05:08 PM, Alain.Moulle wrote: > >> I don't think it is authorized at all, we must never "write" directly >> in cib.xml , I tried for other needs and it systematically disturbs >> Pacemaker a lot ! >> We always have to go through some crm commands or similar >> (crm_attributes, etc.) >> but they are not taken in account before the 60s are ended. >> >> Alain >> >> Dominik Klein a écrit : >> >>> Just write it to the xml on all nodes? >>> >>> On 05/10/2011 01:23 PM, Alain.Moulle wrote: >>> >>> Sorry I meant "directly with is_managed=false" of course ! Alain >>> ___ >>> Linux-HA mailing list >>> Linux-HA@lists.linux-ha.org >>> http://lists.linux-ha.org/mailman/listinfo/linux-ha >>> See also: http://linux-ha.org/ReportingProblems >>> >>> >>> >>> >> ___ >> Linux-HA mailing list >> Linux-HA@lists.linux-ha.org >> http://lists.linux-ha.org/mailman/listinfo/linux-ha >> See also: http://linux-ha.org/ReportingProblems >> >> > > ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
Re: [Linux-HA] How to start Pacemaker in unmanaged mode ?
Correct me if I'm wrong but I strongly think the following would work: Since you want to _start_ pacemaker in unmanaged mode, I expect all your nodes to be offline. Then delete cib.xml on all nodes but one. On the one remaining, edit cib.xml and put your configuration is_managed=false there. Then start all nodes. Don't see why this shouldn't work. Regards Dominik On 05/10/2011 05:08 PM, Alain.Moulle wrote: > I don't think it is authorized at all, we must never "write" directly > in cib.xml , I tried for other needs and it systematically disturbs > Pacemaker a lot ! > We always have to go through some crm commands or similar > (crm_attributes, etc.) > but they are not taken in account before the 60s are ended. > > Alain > > Dominik Klein a écrit : >> Just write it to the xml on all nodes? >> >> On 05/10/2011 01:23 PM, Alain.Moulle wrote: >> >>> Sorry I meant "directly with is_managed=false" of course ! >>> Alain >>> >> ___ >> Linux-HA mailing list >> Linux-HA@lists.linux-ha.org >> http://lists.linux-ha.org/mailman/listinfo/linux-ha >> See also: http://linux-ha.org/ReportingProblems >> >> >> > > ___ > Linux-HA mailing list > Linux-HA@lists.linux-ha.org > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems > -- IN-telegence GmbH Oskar-Jäger-Str. 125 50825 Köln Registergericht AG Köln - HRB 34038 USt-ID DE210882245 Geschäftsführende Gesellschafter: Christian Plätke und Holger Jansen ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
Re: [Linux-HA] How to start Pacemaker in unmanaged mode ?
I don't think it is authorized at all, we must never "write" directly in cib.xml , I tried for other needs and it systematically disturbs Pacemaker a lot ! We always have to go through some crm commands or similar (crm_attributes, etc.) but they are not taken in account before the 60s are ended. Alain Dominik Klein a écrit : > Just write it to the xml on all nodes? > > On 05/10/2011 01:23 PM, Alain.Moulle wrote: > >> Sorry I meant "directly with is_managed=false" of course ! >> Alain >> > ___ > Linux-HA mailing list > Linux-HA@lists.linux-ha.org > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems > > > ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
Re: [Linux-HA] How to start Pacemaker in unmanaged mode ?
Just write it to the xml on all nodes? On 05/10/2011 01:23 PM, Alain.Moulle wrote: > Sorry I meant "directly with is_managed=false" of course ! > Alain ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
Re: [Linux-HA] How to start Pacemaker in unmanaged mode ?
Sorry I meant "directly with is_managed=false" of course ! Alain > Hi > > I wonder if there is a way to start Pacemaker but directly in the mode > is_managed=true, > so that at the start , none action on any resource configured in > Pacemaker will be executed (start, stop, monitoring etc.) > > This would useful in the case we have a big problem which makes all > resources failed in Pacemaker, > and the nodes are rebooted, and we want to start pacemaker again, and do > a crm cleanup on all > resources "failed" before any action of Pacemaker itself. > > In fact, it would be nice to be able to modify any cluster parameter > such as is_managed or others > in the 60s Pacemaker takes to elect the DC when it is the first start of > Pacemaker on the nodes of the cluster. > > Alain > ___ > Linux-HA mailing list > Linux-HA@lists.linux-ha.org > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems > > > ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
[Linux-HA] How to start Pacemaker in unmanaged mode ?
Hi I wonder if there is a way to start Pacemaker but directly in the mode is_managed=true, so that at the start , none action on any resource configured in Pacemaker will be executed (start, stop, monitoring etc.) This would useful in the case we have a big problem which makes all resources failed in Pacemaker, and the nodes are rebooted, and we want to start pacemaker again, and do a crm cleanup on all resources "failed" before any action of Pacemaker itself. In fact, it would be nice to be able to modify any cluster parameter such as is_managed or others in the 60s Pacemaker takes to elect the DC when it is the first start of Pacemaker on the nodes of the cluster. Alain ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
