> ADSM Development does indeed understand that the current database
> dump/load and audit process is waaaay to long running to be
> practical when fast recovery of the server is necessary.
> Informational APAR II08040 describes enhancements that we recently
> introduced to produce a consistent database dump image, and thereby
> completely eliminate the need for a database audit (AUDITDB) after the
> the database is reloaded. In fact, you could have used that method
> to eliminate the AUDITDB that you are currently running.
> The on-line DUMP DB CONSISTENT=YES command can be used to produce
> a consistent database dump so that an AUDITDB is not required
> after the server database is reloaded from the dump.
> In addition, development is actively working on solutions that
> will improve server recovery time overall.
> Mike Kaczmarski
> ADSM Server Development
Hi Mike,
I have a couple of items now. Our databases are well over 1,000,000 pages
now. I don't even want to think of how long an audit would take now.
Is there a status on the backgrounding of audits yet? I know you mentioned
that this was being talked about (audits running while the server is up).
Also, when we do initial dumps it often overwhelms the server and he is
not able to migrate from disk to tape fast enough (if there are many
small objects). I would like to see the following priorities set:
RESTORE: the highest priority
MIGRATION: second only to restore
BACKUP: most backups are batch jobs so this can be lower
EXPIRATION: the lowest priority
DEL DATA: same as expiration
RECLAIM: same as expiration and del data
We are expecting some 9GB filesystems in the near future and it's
scary to think about the first backup and the first full restore!
One server we have is displaying the following characteristics:
VM server, RS/6000 client (model 30), 165 nfs mounted filesystems
from IBM Raven and Auspex (approximately 400GB, 2 million objects
under management, with some exclusions)
adsm> q db
Available Assigned Maximum Maximum Page Total Used %Util Max.
Space Capacity Extension Reduction Size Usable Pages %Util
(MB) (MB) (MB) (MB) (bytes) Pages
--------- -------- --------- --------- ------- --------- --------- ----- -----
7,020 7,020 0 2,612 4,096 1,797,120 1,127,982 62.8 62.8
7,020 7,020 0 2,612 4,096 1,797,120 1,127,982 62.8 62.8
currently (this minute):
adsm> q stg
Storage Device Estimated %Util %Migr High Low Next
Pool Name Class Name Capacity Mig% Mig% Storage
(MB) Pool
----------- ---------- ---------- ----- ----- ---- ---- -----------
BACKUPPOOL DISK 29,627.5 82.5 82.3 5 0 BACKUPTAPE
BACKUPPOOL DISK 29,627.5 82.5 82.3 5 0 BACKUPTAPE
BACKUPTAPE CARTRIDGE 596,767.5 64.8 93.6 100 0
(notice we have 29GB of disk at 82.5% capacity!)
adsm> q proc
Process Process Description Status
Number
-------- -------------------- -------------------------------------------------
2 Migration Disk Storage Pool BACKUPPOOL, Moved Files: 31989,
2 Migration Disk Storage Pool BACKUPPOOL, Moved Files: 31989,
Moved Bytes: 887,197,696, Unreadable Files: 0,
Unreadable Bytes: 0. Current File (bytes): None
Current output volume: DS1467.
3 Expiration ANR0817I Currently processing filespace
/home/aix/p134 for Node DCRST2: have deleted a
total of 35784 backup files, and 0 archive
files.
I restarted the server about 18 hours ago (due to another problem) and you
can see that it moved less than a 1 GB and about 32,000 objects so far.
adsm> q node
Node Name Platform Policy Domain Days Since Days Since Locked?
Name Last Password
Access Set
------------------------- -------- -------------- ---------- ---------- -------
DCRST2 AIX STANDARD <1 224 No
DCRST2 AIX STANDARD <1 224 No
This might appear strange but we only have one node on this server but it
backs up five large fileservers
adsm> q ses
Sess Comm. Sess Wait Bytes Bytes Sess Platform Client Name
Number Method State Time Sent Recvd Type
------ ------ ------ ------ ------- ------- ----- -------- --------------------
73 Tcp/Ip Run 0 S 4.6 M 419.9 M Node AIX DCRST2
73 Tcp/Ip Run 0 S 4.6 M 419.9 M Node AIX DCRST2
100 Tcp/Ip RecvW 6 S 344.8 K 349.9 M Node AIX DCRST2
151 Tcp/Ip Run 0 S 2.3 K 228 Admin SunOS MES00
152 Tcp/Ip Run 0 S 234.5 K 192.5 K Node AIX DCRST2
Currently we have only three sessions running but it can get up to 20 or more
at times, depending on the amount of data changed. I am continuously
firing off backup jobs every 8 minutes (I am tuning that interval now
but 8 minutes fits all filesystems into a 24 hour window).
This is an extreme case but it illustrates the pressure that our server
is under. We are using a Storage Technology Silo and operation is
totally lights-out, however a prioritized migration would help us
greatly. Thanks, Mike. Good to hear from you again.
Mitch
________________________________________________________________________
| Mitch Sako \\\\ \\\\ \\\\\\\\ |||||||| /////// // // |
| I-NET Corporation \\ \\ \\ \\ \\ || // // // |
| LSI Logic Contract \\ \\\ \\ \\ || // //////// |
| Mailstop E-195 \\ \\ \\ || // // // |
| 1501 McCarthy Blvd. \\ \\ \\\\\\\\ || /////// // // |
| Milpitas CA 95035 ___o |
| local: mitch@asic _'\ <_ Phone: (408) 433-4187 |
| internet: msako AT lsil DOT com (_)/ (_) FAX: (408) 433-8796 |
| personal: msako AT netcom DOT com Pager: (408) 989-3365 |
| ibmmail: USMILUN9 AT IBMMAIL |
| DISCLAIMER: Opinions expressed are mine alone |
|______________________________________________________________________|
|