Dear Managers,
i rebooted first member2, than member3 (that was the member where cluster_lockd
was relocated) than everything seems to be fine - except i have still problems
with one bus on a double-SCSI-controller.
6 disks on one bus of the double-controller reports problems - i connected them
to another controller (slower SCSI).
I suggest the controller is defective after power failure 2 weeks ago.
Thanks to Danielle Georgette which suggests the prpasswd is the origin of the
problem - it was not. And william_f_fiore_jr who told me:
I recently had a similar situ w/ lockd disappearing...but system still running and unusable.
I shutdown all the cluster nodes (I had to 'crash dump' the nodes w/ control-P).
Then none of the nodes would boot up.....It turned out that a disk controller was in a 1/2 on - 1/2 off
state....reseting the disk controller by plugging a VT100 into it.....fixed it.
I would have removed power from the systems (which would have cleared the 1/2on-off state) but
the system had been powered on for more than 8 years.....I was afraid that the disks would not spin back up....
so It took 1/2 a day to find and reset the disk controllar with the 1/2on-off situ
Thanks, Harald Baumgartner
--
Harald Baumgartner Max-Planck-Inst. fuer Extraterrestrische Physik
Postfach 1312, D-85741, Garching bei Muenchen, Germany
Phone :(Country Code 49) 89 30000-3346 or -4346
e-mail: hmb_at_mpe.mpg.de (Fax: 49-89/30000-3569)
http://www.xray.mpe.mpg.de/~hmb/
Received on Mon Jun 27 2005 - 15:30:50 NZST