Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
Hi Lars, yes now is more clar for me. We made "node standby" to simulate a package switching ( in a wrong way it seems ). Let me ask one thing: do ypu know a way to do a "cluster functionalty" test? I'm thinking about unplugging lan cable... is that a right test? Thanks again for your support. Simone 2013/2/26 Lars Ellenberg <lars.ellenberg at linbit.com> > On Tue, Feb 26, 2013 at 10:31:27AM +0100, Simone Del Pinto wrote: > > Thank u lars! > > > > I've implemented a cron scripts that if found drbd in Unconfigured state > do > > a service reload. With this workaruond all seems to work fine. > > Then you don't understand your cluster manager at all. > Get rid of this cron job as soon as possible. > See below. > > > But... if the service really switch on other node and, for a reason, it > > come back to last nod what happen? > > > > It is very strange... it need a manual restart however. > > I don't really understand your question. > > But if you want some node to be able to take over, > you must not set it "standby". > > Maybe some misundernstanding on the concept of "node standby" in pacemaker? > > In pacemaker you switch a node to "standby" to take it down for > maintenance. > > In normal operation of the cluster, all nodes must be *online*. > Some of the online nodes run resources, some run other resources, > some run instances of clones of the same resource, > some run instances of multi-state resources as Master or Slave. > > The node state "standby" in the Pacemaker sense is one that does not > really take part in clustering anymore, possibly because it is about > to be taken down for maintenance/hardware upgrade/whatever. > > What you apparently think of "standby" would be a node waiting to take > over. For that to happen, it must be "online" in the pacemaker sense, > just not running the resources in question yet (but probably the Slave > instances of the DRBDs). > > Hope that clears things up a bit. > > Lars > > > > > Simone > > > > 2013/2/26 Lars Ellenberg <lars.ellenberg at linbit.com> > > > > > On Fri, Feb 22, 2013 at 11:26:07AM +0100, Simone Del Pinto wrote: > > > > Hi guys, > > > > > > > > we are using drbd 8.3.13 on our 2 linux server to keep data of the > oer > > > MySQL server. > > > > > > > > corosync and pacemaker ensure that we have a virtuel IP running and > that > > > our DB is always up on one of those nodes. > > > > During a failover test we noted that when we put the Master node in > > > standy > > > > ( crm node standby ) > > > > > > "crm node standby" means: stop all services on that node. > > > stop means stop. > > > stop of DRBD means taking it down. > > > > > > there is no "native" way to > > > tell pacemaker to "go standby but leave DRBD configured as Secondary". > > > > > > There are a few ways to achieve similar functionality, though. > > > You could do implement that using constraints, > > > possibly referencing node attributes. > > > > > > > > > Or you re-configure drbd by hand > > > once you switched to standby in pacemaker, > > > just as you apparently did already. > > > > > > > all services switch to other node but DRBD goes in "Unconfigured > state". > > > > Below the situation before standby: > > > > > > > > crm_mon -V1 > > > > Master/Slave Set: ms_zends_drbd [p_zends_drbd] > > > > Masters: [ FRCVD2046 ] > > > > Slaves: [ FRCVD2047 ] > > > > > > > > Now the situation after "crm node standby": > > > > > > > > Master/Slave Set: ms_zends_drbd [p_zends_drbd] > > > > Masters: [ FRCVD2047 ] > > > > Stopped: [ p_zends_drbd:1 ] > > > > > > > > And now DRBD situation on "old" Master node ( FRCVD2046 ): > > > > > > > > service drbd status > > > > 0:zendsdata Unconfigured > > > > > > Right. > > > Works "as designed". > > > > > > > Now if i run a DRBD service reload I have this scenario: > > > > > > > > service drbd reload > > > > service drbd status > > > > 0:zendsdata Connected Secondary/Primary UpToDate/UpToDate C > > > > > > Of course. > > > > > > > > > -- > > > : Lars Ellenberg > > > : LINBIT | Your Way to High Availability > > > : DRBD/HA support and consulting http://www.linbit.com > > > > > > DRBD® and LINBIT® are registered trademarks of LINBIT, Austria. > > > _______________________________________________ > > > drbd-user mailing list > > > drbd-user at lists.linbit.com > > > http://lists.linbit.com/mailman/listinfo/drbd-user > > > > > > _______________________________________________ > > drbd-user mailing list > > drbd-user at lists.linbit.com > > http://lists.linbit.com/mailman/listinfo/drbd-user > > > -- > : Lars Ellenberg > : LINBIT | Your Way to High Availability > : DRBD/HA support and consulting http://www.linbit.com > > DRBD® and LINBIT® are registered trademarks of LINBIT, Austria. > __ > please don't Cc me, but send to list -- I'm subscribed > -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.linbit.com/pipermail/drbd-user/attachments/20130226/e0579d05/attachment.htm>