On 12 Nov 2013, at 7:44 am, Tom Parker <tpar...@cbnco.com> wrote: > I have found my settings. > > I needed to set the following in /etc/sysconfig/pacemaker > > # Force use of a particular class of IPC connection > > # PCMK_ipc_type=shared-mem|socket|posix|sysv > > export PCMK_ipc_type=shared-mem
This is the default now btw. > > # Specify an IPC buffer size in bytes > > # Useful when connecting to really big clusters that exceed the default 20k > buffer > > # PCMK_ipc_buffer=20480 > > export PCMK_ipc_buffer=20480000 20Mb is pretty big, unless you're wanting to avoid compression completely. If you run: cibadmin -Ql > big.xml bzip big.xml ls -alh big.xml.bz2 and then round up a bit more, you should be pretty safe. > > > and in my bashrc file (for the crm tools to work proplery) I have > > [LIVE] qaxen1:~ # cat .bashrc > > # Load Pacemaker IPC settings for crm > > PACEMAKER_SYSCONFIG=/etc/sysconfig/pacemaker > > if [ -f $PACEMAKER_SYSCONFIG ]; then > > . $PACEMAKER_SYSCONFIG > > fi This should be redundant in the upcoming 1.1.11 release. David is working on a patch that will allow the IPC clients to pick up the larger default automagically. > > > Hope this helps. > > On 11/11/2013 03:35 PM, Tom Parker wrote: >> You will also have to be careful of the shared memory size between the >> nodes. I had issues with massive cibs. Setting some environment variables >> fixed the issue but the defaults are too small. >> >> From: Digimer >> Sent: Monday, November 11, 2013 10:24 AM >> To: General Linux-HA mailing list >> Reply To: General Linux-HA mailing list >> Subject: Re: [Linux-HA] How many primitives, groups can I have >> >> >> On 11/11/13 07:57, Michael Brookhuis wrote: >>> Hi, >>> >>> Is there a limit in the number of proimitives, etc you can have? >>> What maximum number is recommended based on best-practices? >>> >>> Are 1500 to many? >>> >>> Thanks >>> Mima >> The cib will be very large, so pushing changes to other nodes will take >> time (specially if you have many nodes). I suspect you will run into >> corosync timeouts before you hit any coded upper limits. You will likely >> have to play with corosync timing values to get that high, assuming your >> network is fast enough at all. >> >> But in the end, as I understand it, there is no coded upper limit. >> >> -- >> Digimer >> Papers and Projects: https://alteeve.ca/w/ >> What if the cure for cancer is trapped in the mind of a person without >> access to education? >> _______________________________________________ >> 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 _______________________________________________ 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