|
From: Thomas Guyot-S. <Th...@za...> - 2006-03-10 14:28:13
|
# evmsccm -v CCM has a membership. The membership has quorum. If you get something different, check you heartbeat configuration. Make sure all nodes defined in ha.cf went up OR have been stonished. I know Linux-HA documentaion isn't well organized, but if you dig enough you'll find answers to all your questions. Good luck! Thomas > -----Original Message----- > From: Eric [mailto:eri...@gm...] > Sent: March 10, 2006 8:30 > To: Thomas Guyot-Sionnest > Cc: evm...@li... > Subject: Re: [Evms-devel] How to use EVMS and Linux-HA together? > > Thank your very much! I have added these two lines into my ha.cf file: > apiaauth evms uid=root > apiaauth ccm uid=hacluster > > Then, there is no longer that alert message when I enter EVMS > interface. > > But now, I still cannot do any cluster related thing, option > Setting -> Node Administrated... is still disabled. I looked > up /var/log/evms-engine.log, and found these error messages: > > _3_ 4029e080 Linux-HA: get_membership: Error: Membership not > initialized yet. Try again. > ... > _3_ 4029e080 do_fork_and_execvp: execvp() failed. errno is 2: > No such file or directory. > _3_ 4029e080 do_fork_and_execvp: execvp() failed. errno is 2: > No such file or directory. > ... > _3_ 4029e080 do_fork_and_execvp: execvp() failed. errno is 2: > No such file or directory. > _3_ 4029e080 do_fork_and_execvp: execvp() failed. errno is 2: > No such file or directory. > ... > _3_ 4029e080 Engine: engine_open_object: Open of > /dev/evms/.node/hdc failed with error code > 123: No medium found. > _3_ 4029e080 Engine: engine_open_object: Open of > /dev/evms/.node/ram0 failed with error code 22: Invalid argument. > _3_ 4029e080 Engine: engine_open_object: Open of > /dev/evms/.node/ram1 failed with error code 22: Invalid argument. > _3_ 4029e080 Engine: engine_open_object: Open of > /dev/evms/.node/ram10 failed with error code 22: Invalid argument. > ... > > I don't know why these messages appeared. > By the way, I am using evms only in one machine(node), should > I set up a real cluster(contains two or more node) by heartbeat first? > > Hope for your reply! > > > > Thanks and Regards, > Eric > > |