Re: [Pacemaker] fencing best practices for virtual environments

Hi Lars, thank you very much for the deep explanation. Regards, Alberto On 09/10/2012 03:42 PM, Lars Marowsky-Bree wrote: On 2012-09-10T14:40:43, Alberto Menichetti wrote: Sorry, maybe I'm missing something, but suppose this scenario (also remember that, being a 2-node cluster, I had to set

Re: [Pacemaker] CRM CLI guide gone?

On Tue, Sep 11, 2012 at 02:17:56PM +0200, Andreas Ntaflos wrote: > On 11/09/12 13:48, Dejan Muhamedagic wrote: > >> Is it perhaps ? > > > > Yes. > > Ok, great to know, thanks! > > > I'm a bit low on html, sorry about that. But something will be > > pos

Re: [Pacemaker] CRM CLI guide gone?

On 11/09/12 13:48, Dejan Muhamedagic wrote: >> Is it perhaps ? > > Yes. Ok, great to know, thanks! > I'm a bit low on html, sorry about that. But something will be > posted soon. At any rate, the crm(8) man page should be available. At least on Ubuntu

Re: [Pacemaker] CRM CLI guide gone?

Hi, On Mon, Sep 10, 2012 at 06:26:34PM +0200, Andreas Ntaflos wrote: > On 10/09/12 00:15, Andrew Beekhof wrote: > > On Fri, Sep 7, 2012 at 6:49 AM, Andreas Ntaflos > > wrote: > >> It seems the CRM CLI guide formally found under > >> > >> http://www.clusterlabs.org/doc/crm_cli.html > >> > >> is g

Re: [Pacemaker] Clustering of Two virtual machines hosted on different physical servers

On Fri, Sep 7, 2012 at 1:25 PM, ecfgijn wrote: > Hi Michael , > > If you don't mind , can you please share the name centos-6.2 repositories > where i can get dlm and kernal-module-extra rpms. As i have already tried > epel and rpmforge and scientific linux and default centos repo. http://www.c

Re: [Pacemaker] [Problem] About the replacement of the master/slave resource.

On Mon, Sep 10, 2012 at 4:42 PM, wrote: > Hi All, > > We confirmed movement of the trouble of the clone resource that we combined > with Master/Slave resource. > > The master / slave resources are replaced under the influence of the trouble > of the clonal resource. > > We confirmed it in the n

Re: [Pacemaker] master/slave resource does not stop (tries start repeatedly)

On Tue, Sep 11, 2012 at 9:13 PM, Andrew Beekhof wrote: > Yikes! > > Fixed in: >https://github.com/beekhof/pacemaker/commit/7d098ce That link should have been: https://github.com/beekhof/pacemaker/commit/c1f409baaaf388d03f6124ec0d9da440445c4a23 > > On Fri, Sep 7, 2012 at 7:49 PM, Kazuno

Re: [Pacemaker] Getting Error while Active/Active clustering in centos-6.2

Follow the instructions in the edition of "Clusters from Scratch" for your software versions. Particularly Chapter 8: http://www.clusterlabs.org/doc/en-US/Pacemaker/1.1-plugin/html/Clusters_from_Scratch/ch08.html On Fri, Sep 7, 2012 at 4:34 PM, ecfgijn wrote: > > Hi All , > > I have followed

Re: [Pacemaker] master/slave resource does not stop (tries start repeatedly)

Yikes! Fixed in: https://github.com/beekhof/pacemaker/commit/7d098ce On Fri, Sep 7, 2012 at 7:49 PM, Kazunori INOUE wrote: > Hi, > > I am using Pacemaker-1.1. > - ClusterLabs/pacemaker : 872a2f1af1 (Sep 07) > > Though a monitor of master resource fails and there is no node which > the master/

[Pacemaker] VirtualDomain shutdown problem

Hi. I've an installation with a VirtualDomain KVM machine which doesn't support shutdown, so it has to be destroyed. >From what I see VirtualDomain should first issue a shutdown and after the >timeout destroy the vm, but I'm using Pacemaker 1.1.6 and I found this thread http://www.gossamer-thre

[Pacemaker] Bandwidth needs of corosync

Hi, Are there any figures of the need of bandwidth of corosync? Does it scale with the number of nodes? Greetings, -- Dr. Michael Schwartzkopff Guardinistr. 63 81375 München Tel: (0163) 172 50 98 Fax: (089) 620 304 13 signature.asc Description: This is a digitally signed message part. _

Re: [Pacemaker] using op_defaults to change default operation failure action?

On 11/09/2012 10:37, Lars Marowsky-Bree wrote: On 2012-09-11T10:24:58, Nenad Opsenica wrote: That was my backup solution, but could not find any working example of how to do this using crm shell. Do you have some pointers or examples? The crm online help on rsc_template? http://hg.savannah.g

Re: [Pacemaker] ordering such that at least one resource started

Hi, On Tue, Sep 11, 2012 at 05:57:41PM +1000, Andrew Beekhof wrote: > The feature will be released for the first time in 1.1.8 > The shell might not support it yet, you might need to use the xml until then The shell is up to date now. This would be the syntax: order dc_then_exch 0: [ dc1 dc2 ] e

Re: [Pacemaker] using op_defaults to change default operation failure action?

On 2012-09-11T10:24:58, Nenad Opsenica wrote: > That was my backup solution, but could not find any working example > of how to do this using crm shell. > Do you have some pointers or examples? The crm online help on rsc_template? http://hg.savannah.gnu.org/hgweb/crmsh/file/bb3dabeb8274/doc/crm

Re: [Pacemaker] CRM CLI guide gone?

On 11/09/2012 10:11, Rasto Levrinc wrote: The crm shell is the "official" CLI way. If you are interested in GUIs, there are several, but LCMC that has still the most features until the elephants come up with something better, which they IMHO didn't.. :) Rasto Yup, I'm using LCMC also, excelle

Re: [Pacemaker] using op_defaults to change default operation failure action?

On 11/09/2012 10:09, Lars Marowsky-Bree wrote: On 2012-09-11T10:02:21, Nenad Opsenica wrote: Thank you Andrew. Now it is more clear to me. Anyhow, it definitely would be nice to have a option to specify actions per-operation ... Please account one vote from me for this feature ;) You can use

Re: [Pacemaker] pacemaker processes RSS growth

07.09.2012 09:25, Vladislav Bogdanov wrote: > 06.09.2012 12:58, Vladislav Bogdanov wrote: > ... >> lrmd seems not to clean up gio channels properly: > > I prefer to call g_io_channel_unref() right after g_io_add_watch_full() > instead of doing so when deleting descriptor (g_source_remove() is > en

Re: [Pacemaker] CRM CLI guide gone?

On Mon, Sep 10, 2012 at 10:15 PM, Nenad Opsenica wrote: > > On 09/10/2012 06:26 PM, Andreas Ntaflos wrote: >> >> On 10/09/12 00:15, Andrew Beekhof wrote: >>> >>> On Fri, Sep 7, 2012 at 6:49 AM, Andreas Ntaflos >>> wrote: It seems the CRM CLI guide formally found under http://w

Re: [Pacemaker] using op_defaults to change default operation failure action?

On 2012-09-11T10:02:21, Nenad Opsenica wrote: > Thank you Andrew. Now it is more clear to me. > Anyhow, it definitely would be nice to have a option to specify > actions per-operation ... Please account one vote from me for this > feature ;) You can use the template resource feature and then inh

Re: [Pacemaker] using op_defaults to change default operation failure action?

On 11/09/2012 09:55, Andrew Beekhof wrote: On Mon, Sep 10, 2012 at 4:55 AM, Nenad Opsenica wrote: Hi everybody, Is it possible to use "op_defaults" to specify default failure action for specific operation? No. The values are used as the default for all operations. I'm particulary interested

Re: [Pacemaker] ordering such that at least one resource started

The feature will be released for the first time in 1.1.8 The shell might not support it yet, you might need to use the xml until then On Fri, Sep 7, 2012 at 12:07 PM, James Harper wrote: > Further to my last email, I now want to create an ordering rule that says > "exchange server should only st

Re: [Pacemaker] using op_defaults to change default operation failure action?

On Mon, Sep 10, 2012 at 4:55 AM, Nenad Opsenica wrote: > Hi everybody, > > Is it possible to use "op_defaults" to specify default failure action for > specific operation? No. The values are used as the default for all operations. > > I'm particulary interested in specifying action if "stop" fail

Re: [Pacemaker] Possible bug?

On Tue, Sep 11, 2012 at 4:58 PM, Borislav Borisov wrote: > >> Pacemaker creates a second cib process to write the contents to disk >> after a change so that the 'real' process doesn't block. >> >> After the process writes the cib to disk, we then try to read it back >> again to verify that everyth

Re: [Pacemaker] Possible bug?

> Pacemaker creates a second cib process to write the contents to disk > after a change so that the 'real' process doesn't block. > > After the process writes the cib to disk, we then try to read it back > again to verify that everything is sane. > What you're seeing here is that check failing for