I see.
You can remove the dir from the devclass, but TSM will still access existing
volumes on that dir. And then a procedure to movedata them.
My assumption was that you wouldn't be able to remove the dir from the devclass
until while there were vols on it, and I couldn't think of how to get rid of
the vols on the dir since it was in use.
Great Info!
Thanks!
Rick
-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On Behalf Of
Thomas Denier
Sent: Tuesday, February 14, 2017 11:47 AM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: *EXTERNAL* Re: Multiple NFS mounts to same DataDomain
I have successfully removed directories from file device classes a number of
times. The sequence of events was as follows:
1.Remove the directory from the list in the device class definition to prevent
allocation of new scratch volumes on the directory.
2.List filling volumes on the directory and make them read only, preventing
appends to existing volumes on the directory.
3.Run "move data" commands to relocate the contents of volumes to other
directories.
4.Wait for database backups and pending volumes to age off.
The Linux administrator thought step 3 would have better overall throughput
with multiple streams of "move data" commands. I used the "-P" option of
"xargs" to manage the multiple streams.
Thomas Denier,
Thomas Jefferson University
-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On Behalf Of
Rhodes, Richard L.
Sent: Tuesday, February 14, 2017 08:35
To: ADSM-L AT VM.MARIST DOT EDU
Subject: [ADSM-L] Multiple NFS mounts to same DataDomain
Arnaud's discussion on the another thread is SO interesting (Availability for
Spectrum Protect 8.1 server software for Linux on power system).
It got me thinking of our problems . . .
> NFS, whose performance is not that good on AIX systems
Agreed!!! After getting DataDomain system and using NFS we were/are VERY
unhappy with the NFS performance.
Our Unix admins worked with IBM/AIX support, and finally got an admission that
the problem is AIX/NFS using a single TCP socket for all writes. The
workaround was to use multiple mount point to the same NFS share and spread
writes (somehow) across them. He did this and got higher throughput.
So now I'm wondering if we could use multiple NFS mounts to the same DD for our
file device pools.
aix: /DD/tsm1/mnt1 dd: /data/col1/tsm1/mnt1
/DD/tsm1/mnt2 /data/col1/tsm1/mnt2
/DD/tsm1/mnt3 /data/col1/tsm1/mnt2
Then use multiple dir's for the file device devclass:
define devclass DDFILEDEV devtype=file
dir=/DD/tsm1/mnt1,/DD/tsm1/mnt2,/DD/tsm1/mnt3
According to the link to dsmISI again, TSM will roughly balance across the
multiple mount points, hopefully giving better write throughput. I've been
VERY reluctant to try this since it appears once you add a dir to a file device
devclass, it's there forever!
I'm curious if anyone is doing this.
Rick
The information contained in this transmission contains privileged and
confidential information. It is intended only for the use of the person named
above. If you are not the intended recipient, you are hereby notified that any
review, dissemination, distribution or duplication of this communication is
strictly prohibited. If you are not the intended recipient, please contact the
sender by reply email and destroy all copies of the original message.
CAUTION: Intended recipients should NOT use email communication for emergent or
urgent health care matters.
-----------------------------------------
The information contained in this message is intended only for the personal and
confidential use of the recipient(s) named above. If the reader of this message
is not the intended recipient or an agent responsible for delivering it to the
intended recipient, you are hereby notified that you have received this
document in error and that any review, dissemination, distribution, or copying
of this message is strictly prohibited. If you have received this communication
in error, please notify us immediately, and delete the original message.
|
