On Fri, Feb 15, 2008 at 2:52 AM, Luis Martin-Santos <[EMAIL PROTECTED]> wrote: > ... ¿Have you configured STONITH?
I followed 2 different tutorials simultaneously as each one was missing key information and NEITHER one mentioned STONITH or Fencing at all. In the case of the tutorial I found in the blog that no one commented that the tutorial would lead to a split-brain situation. So to answer your question I have not configured STONITH. On Fri, Feb 15, 2008 at 4:57 AM, <[EMAIL PROTECTED]> wrote: > Split-Brain Situations are *very* critical for a two node setup, aspecially > when you are using > shared media like disks drbd syncs and so on. > > For bigger clusters the problem is a bit more easy, bevause you get a quorum > loss, if half the > nodes are "down" or disconnected. You can use the directive for the no_qurum > (ignore, *stop*, fence) > to make that stable - the ressources are stopped, ehen you use the value stop > for that directive. > > For a two node setup you run the twonode quorum plugin, which does not stop > the ressources in the > case of a split (if I got that correctly teh quorum is not lost). Thus both > nodes *could* aquire the resources. Funny how none of these were mentioned in the tutorials. Hopefully I will have the time to prepare a tutorial that *will* cover fencing and such but I am still just beginning the learning curve. I did a quick look around on linux-ha and I have to admit the info that is there is hard to find effectively. I looked at my system install and I see that there are shared libraries pertaining to plugins so at least my question on where to get them has already been answered. Could you direct me somewhere to find instructions on how to setup this and other plugins I will need to use? Also does this plugin work in a version 1 configuration? > To come out of those problems you need to setup STONITH (server fencing) > resources. In your > setup you would need something like a management board, remote power > switch,... so one node > could switch-off or reboot the other one. Does STONITH work in concert with the quorum plugin or is it something totally different? I know that STONITH rpm was installed when I installed the heartbeat package. By googling I found this http://www.linux-ha.org/ConfiguringStonithPlugins so I will look to that for starters. Is there a better more full blown coverage of the STONITH plugin somewhere? > Once the rebooted node comes up it should not automatically restart > heartbeat, to avoid running into > a second split brain situation. Great! It looks like I have at least a direction to go in so that I can prevent this from happening again. However I am at a total loss as to how to get back to a stable state. The output of /proc/drbd --> version: 8.0.8 (api:86/proto:86) GIT-hash: bd3e2c922f95c4fa0dca57a4f8c24bf8b249cc02 build by [EMAIL PROTECTED], 2008-02-01 07:33:35 0: cs:StandAlone st:Secondary/Unknown ds:UpToDate/DUnknown r--- ns:0 nr:0 dw:8 dr:249 al:0 bm:256 lo:0 pe:0 ua:0 ap:0 resync: used:0/31 hits:0 misses:0 starving:0 dirty:0 changed:0 act_log: used:0/257 hits:2 misses:0 starving:0 dirty:0 changed:0 Looks as if the Primary cannot contact the Secondary but I know the link is alive. How do I get back to a stable state? This is just a test so no data has been lost nor do I care about anything in the shared partition. Thanks regards, Douglas Lochart _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
