Re: [ClusterLabs] DLM standalone without crm ?
On 27/06/16 15:27, Lentes, Bernd wrote: > > - Am 27. Jun 2016 um 9:04 schrieb Christine Caulfield ccaul...@redhat.com: > >> On 26/06/16 14:47, Lentes, Bernd wrote: >>> >>> >>> - Am 26. Jun 2016 um 7:59 schrieb Ferenc Wágner wf...@niif.hu: >>> "Lentes, Bernd"writes: > wf...@niif.hu writes: > >> "Lentes, Bernd" writes: >> >>> is it possible to have a DLM running without CRM? >> >> Yes. You'll need to configure fencing, though, since by default DLM >> will try to use stonithd (from Pacemaker). But DLM fencing didn't >> handle fencing failures correctly for me, resulting in more nodes >> being fenced until quorum was lost, > > do you have a howto ? Or is there one on the net ? dlm-controld has both a SysV init script and a systemd service file, you can use either one to start it (they take care of mounting configfs and loading the necessary kernel modules). Just start it on your nodes when you have quorum. >>> >>> Hi, >>> >>> i don't have neither an init-script nor a systemd service file. >>> The only packages i find in the repositories concerning dlm are: >>> libdlm3-3.00.01-0.31.87 >>> libdlm-3.00.01-0.31.87 >>> And i have a kernel module for dlm. >>> Nothing else. >>> >>> >> >> >> If you have that version of DLM then the cman package will have the >> dlm_controld daemon and init scripts in it. With that version you can >> run DLM without crm but not without cman. cman is required. >> >> Chrissie >> > > Hi Chrissie, > > Isn't cman a crm ? Basically the same as pacemaker ? > It seems in the end i have to use a crm. > So i will try it directly with pacemaker, because i have a SuSE system, > and cman is Red Hat stuff, isn't it ? > > Sorry, I didn't recognise the acronym CRM as a generic - I thought you meant the crmd part of pacemaker. Yes, DLM needs a cluster manager to work. With DLM3 that's usually cman, if you want to do it under pacemaker then you might have to do more work than you'd like. DLM4 should be fine though Chrissie ___ Users mailing list: Users@clusterlabs.org http://clusterlabs.org/mailman/listinfo/users Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org
Re: [ClusterLabs] Pacemaker and Corosync versions compatibility
On 06/24/2016 08:46 PM, Maciej Kopczyński wrote: > Hello, > > I've been following a tutorial to set up a simple HA cluster using > Pacemaker and Corosync on CentOS 6.x while I have noticed that in the > original documentation it is stated that: > > "Since |pcs| has the ability to manage all aspects of the cluster (both > corosync and pacemaker), it requires a specific cluster stack to be in > use: corosync 2.0 or later with votequorum plus Pacemaker 1.1.8 or later." That's incorrect -- pcs has supported RHEL 6 and its derivatives (with corosync 1 and CMAN) since at least 6.5. Thanks for bringing it up; I'll delete that note. > Here are the versions of packages installed in my system (CentOS 6.7): > pacemaker-1.1.14-8.el6.x86_64 > corosync-1.4.7-5.el6.x86_64 > > I did not do that much of testing, but my cluster seems to be more or > less working so far, what are the compatibility issues then? What will > not work with corosync in version lower than 2.0? corosync 2 is a cleaner implementation and has better message delivery guarantees, but most of that is transparent to the user. With corosync 1, pacemaker will use a "legacy" version of the node attribute daemon, so the attrd_updater command has fewer options. Most of the other differences are in implementation, and may be less efficient for example, but are not readily visible. > Thanks in advance for your answers. > > MK ___ Users mailing list: Users@clusterlabs.org http://clusterlabs.org/mailman/listinfo/users Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org
Re: [ClusterLabs] DLM standalone without crm ?
- Am 27. Jun 2016 um 9:04 schrieb Christine Caulfield ccaul...@redhat.com: > On 26/06/16 14:47, Lentes, Bernd wrote: >> >> >> - Am 26. Jun 2016 um 7:59 schrieb Ferenc Wágner wf...@niif.hu: >> >>> "Lentes, Bernd"writes: >>> wf...@niif.hu writes: > "Lentes, Bernd" writes: > >> is it possible to have a DLM running without CRM? > > Yes. You'll need to configure fencing, though, since by default DLM > will try to use stonithd (from Pacemaker). But DLM fencing didn't > handle fencing failures correctly for me, resulting in more nodes > being fenced until quorum was lost, do you have a howto ? Or is there one on the net ? >>> >>> dlm-controld has both a SysV init script and a systemd service file, you >>> can use either one to start it (they take care of mounting configfs and >>> loading the necessary kernel modules). Just start it on your nodes when >>> you have quorum. >> >> Hi, >> >> i don't have neither an init-script nor a systemd service file. >> The only packages i find in the repositories concerning dlm are: >> libdlm3-3.00.01-0.31.87 >> libdlm-3.00.01-0.31.87 >> And i have a kernel module for dlm. >> Nothing else. >> >> > > > If you have that version of DLM then the cman package will have the > dlm_controld daemon and init scripts in it. With that version you can > run DLM without crm but not without cman. cman is required. > > Chrissie > Hi Chrissie, Isn't cman a crm ? Basically the same as pacemaker ? It seems in the end i have to use a crm. So i will try it directly with pacemaker, because i have a SuSE system, and cman is Red Hat stuff, isn't it ? Bernd Helmholtz Zentrum Muenchen Deutsches Forschungszentrum fuer Gesundheit und Umwelt (GmbH) Ingolstaedter Landstr. 1 85764 Neuherberg www.helmholtz-muenchen.de Aufsichtsratsvorsitzende: MinDir'in Baerbel Brumme-Bothe Geschaeftsfuehrer: Prof. Dr. Guenther Wess, Dr. Alfons Enhsen, Renate Schlusen (komm.) Registergericht: Amtsgericht Muenchen HRB 6466 USt-IdNr: DE 129521671 ___ Users mailing list: Users@clusterlabs.org http://clusterlabs.org/mailman/listinfo/users Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org
Re: [ClusterLabs] [corosync] [build] the configure shows UNKNOWN version and runtime exits 17
Hello. I use this guide [0] to build libqb, corosync, pacemaker and test them as pid-space linked docker containers [1]. A Pacemaker builds OK and shows the v1.1.15 runtime, a build-time it complains about an unknown libqb version. I workarounded it by running a Pacemaker build with these env vars set: export PREFIX=/usr export PKG_CONFIG_PATH=${PREFIX}/lib/pkgconfig export libqb_CFLAGS="-I${PREFIX}/include" export libqb_LIBS="-L${PREFIX}/lib" But corosync ./configure shows an unknown version: corosync configuration: Version = UNKNOWN https://github.com/corosync/corosync/issues/116 Use official signed tarbal ( http://build.clusterlabs.org/corosync/releases/corosync-2.3.6.tar.gz) or full git clone. Do not use github "releases". Regards, Honza Although it builds, runtime it throws the exit code 17 on me, which I failed to google / understand from sources. Any help with that, folks? Thank you! [0] http://clusterlabs.org/wiki/SourceInstall [1] https://github.com/docker/docker/blob/master/docs/reference/run.md#pid-settings---pid ___ Users mailing list: Users@clusterlabs.org http://clusterlabs.org/mailman/listinfo/users Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org
Re: [ClusterLabs] [corosync] [build] the configure shows UNKNOWN version and runtime exits 17
On 06/27/2016 09:28 AM, Bogdan Dobrelya wrote: > Hello. > I use this guide [0] to build libqb, corosync, pacemaker > and test them as pid-space linked docker containers [1]. > > A Pacemaker builds OK and shows the v1.1.15 runtime, a build-time it > complains about an unknown libqb version. I workarounded it by running a > Pacemaker build with these env vars set: > > export PREFIX=/usr > export PKG_CONFIG_PATH=${PREFIX}/lib/pkgconfig > export libqb_CFLAGS="-I${PREFIX}/include" > export libqb_LIBS="-L${PREFIX}/lib" > > > But corosync ./configure shows an unknown version: > > corosync configuration: > Version = UNKNOWN > > Although it builds, runtime it throws the exit code 17 on me, which I > failed to google / understand from sources. Any help with that, folks? > Thank you! > > [0] http://clusterlabs.org/wiki/SourceInstall > [1] > https://github.com/docker/docker/blob/master/docs/reference/run.md#pid-settings---pid > Few more, details, with a debug enabled, it shows Jun 27 07:59:15 notice [MAIN ] Corosync Cluster Engine ('UNKNOWN'): started and ready to provide service. Jun 27 07:59:15 info[MAIN ] Corosync built-in features: pie relro bindnow Jun 27 07:59:15 error [MAIN ] Corosync Executive couldn't create lock file. Jun 27 07:59:15 error [MAIN ] Corosync Cluster Engine exiting with status 17 at main.c:1357. The config file I use is http://pastebin.com/ArBBtQTy -- Best regards, Bogdan Dobrelya, Irc #bogdando ___ Users mailing list: Users@clusterlabs.org http://clusterlabs.org/mailman/listinfo/users Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org
[ClusterLabs] [corosync] [build] the configure shows UNKNOWN version and runtime exits 17
Hello. I use this guide [0] to build libqb, corosync, pacemaker and test them as pid-space linked docker containers [1]. A Pacemaker builds OK and shows the v1.1.15 runtime, a build-time it complains about an unknown libqb version. I workarounded it by running a Pacemaker build with these env vars set: export PREFIX=/usr export PKG_CONFIG_PATH=${PREFIX}/lib/pkgconfig export libqb_CFLAGS="-I${PREFIX}/include" export libqb_LIBS="-L${PREFIX}/lib" But corosync ./configure shows an unknown version: corosync configuration: Version = UNKNOWN Although it builds, runtime it throws the exit code 17 on me, which I failed to google / understand from sources. Any help with that, folks? Thank you! [0] http://clusterlabs.org/wiki/SourceInstall [1] https://github.com/docker/docker/blob/master/docs/reference/run.md#pid-settings---pid -- Best regards, Bogdan Dobrelya, Irc #bogdando ___ Users mailing list: Users@clusterlabs.org http://clusterlabs.org/mailman/listinfo/users Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org
Re: [ClusterLabs] DLM standalone without crm ?
On 26/06/16 14:47, Lentes, Bernd wrote: > > > - Am 26. Jun 2016 um 7:59 schrieb Ferenc Wágner wf...@niif.hu: > >> "Lentes, Bernd"writes: >> >>> wf...@niif.hu writes: >>> "Lentes, Bernd" writes: > is it possible to have a DLM running without CRM? Yes. You'll need to configure fencing, though, since by default DLM will try to use stonithd (from Pacemaker). But DLM fencing didn't handle fencing failures correctly for me, resulting in more nodes being fenced until quorum was lost, >>> >>> do you have a howto ? Or is there one on the net ? >> >> dlm-controld has both a SysV init script and a systemd service file, you >> can use either one to start it (they take care of mounting configfs and >> loading the necessary kernel modules). Just start it on your nodes when >> you have quorum. > > Hi, > > i don't have neither an init-script nor a systemd service file. > The only packages i find in the repositories concerning dlm are: > libdlm3-3.00.01-0.31.87 > libdlm-3.00.01-0.31.87 > And i have a kernel module for dlm. > Nothing else. > > If you have that version of DLM then the cman package will have the dlm_controld daemon and init scripts in it. With that version you can run DLM without crm but not without cman. cman is required. Chrissie ___ Users mailing list: Users@clusterlabs.org http://clusterlabs.org/mailman/listinfo/users Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org