database locking query

pavanvr — Wed, 11 Jul 2007 07:00:54 GMT

this is a situation that I am facing. Hence require urgent help.
two systems S1 and S2. Both of them access a raw
disk RD where oracle tables are present and both systems use their own
tablespaces. One of them is able to access the disk while the other is
waiting (according to the oracle guys). I ran glance and one of the
systems is showing 100% DISK utilization when the batch process is run.
The runq increases and the avwait increases and ultimately the service
time increases.
Is there a problem like when the same raw disk is used by Oracle,
there is some kind of locking mechanism which prevents the other system
to access it? I mean is there a configuration where that locking could
be done? Otherwise parallel access should be possible because they are
using different tablespaces.
Please let me know how I can proceed to debug such problem.

Re: database locking query

Steven E. Protter — Wed, 11 Jul 2007 07:20:08 GMT

Shalom,

If a volume group (disk) is on shared storage the volume group must be activated in shared mode for two database instances to access it from two different machines.

Note that Oracle Server is not certified for this type of access. You would need RAC to successfully do this.

LVM on HP-UX prevents access to the same disk (volume group) from two different systems unless the volume group is activated in shared mode. So, yes there is a locking mechanism and turning it off may lead to data corruption.

SEP

topic database locking query in Operating System - HP-UX

database locking query

Re: database locking query