On 9/17/2012 1:52 PM, Mike B wrote:
Hi,
I'm somewhat new to the storage game. I noticed that our RMM cds is
showing at 84% utilization.
Just so I don't get ambushed with any EDG2109I MASTER FILE IS FULL
warnings, I want to be pre-prepared with reorg jcl should we need to
respond.
The RMM Primer details nicely an in-place reorg using
PGM=EDGBKUP,PARM='BACKUP(REORG)'
It also states that one can bypass the need for taking down the RMM
started task and bypass the need for
taking the tape-drives offline if one simply issues a quiesce command:
F DFRMM,QUIESCE
Question : What happens to any batch jobs that are already running
and using tape at the time that the
F DFRMM,QUIESCE command is issued?
Will RMM wait to implement its quiesce until that job(s) ends? Or is
RMM savvy enough to continue with the quiesce and also smart enough to
update any tape activities appropriately after the RMM reorg completes
and re-inited via F DFRMM,M=xx
We are a very busy shop...there usually is always some job or jobs
using tape around the clock. I fear that if we simply quiesce while a
tape batch job is running, said batch jobs will cause synch problems
or something.
Should I rather decide to go ahead and increase the size of the CDS
and repro it instead of backup-reorg,
I have a question on secondary space. The manuals all have examples
of allocating a cds and it shows
primary and secondary space in cylinders.
I was told that secondary space can be a danger:
We have four lpars, each with their own RMM started tasks, but all
four share the same CDS and journal. We had a prior storage
consultant who was adamant that we should never allocate the CDS with
secondary space...that we must leave the secondary as zero, because,
and I quote:
"In a shared environment, if a the CDS extends on host 1, host 2 would
not know about the extension and if host 2 needs to extend, it would
do its own. "
Is that true?...in a shared environment, we should never code a
secondary space for the CDS?
Lastly, I've already experienced EDG2107E JOURNAL THRESHOLD REACHED
console messages.......Is there any way one can automate the same kind
of warning for when the CDS gets, say 90% utilized?
Mike,
Any tapes going through OCE while RMM is quiesced will get a WTOR to
that effect. After your reorg is complete, you would respond CONTINUE,
or C to those WTOR's (check on that, YMMV). Your RMM CDS should not
typically grow unless the number of datasets written increases, your
tapes get retained longer, and/or you add volumes, VRS's and/or users.
If you have growth issues requiring frequent reorg, you should
investigate why. As far as secondary extents are concerned, the RMM
started tasks should be able to track them, PROVIDED you have the same
CDSID defined to each and they are all actually looking at the same CDS.
If RMM does not track secondary extents on the other systems, you
should open a PMR, because that would be a bug, not a feature. As far
as thresholding, I wrote my own Rexx exec to issue the RMM LC command
and send an Email to the appropriate folks based on a threshold.
Regards,
Tom Conley
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
