ADSM-L

Re: selective vs incremental backup processing times

1999-12-21 03:48:07
Subject: Re: selective vs incremental backup processing times
From: Simon Watson <simon.s.watson AT SHELL.COM DOT BN>
Date: Tue, 21 Dec 1999 16:48:07 +0800
Incremental backup forever is the default.  There are no problems with
this.  A Selective may be useful to consolidate all the active data
into a contiguous set.  This may be helpful if you keep a large number
of versions of files or have a long retention period for deleted files,
as it will significantly reduce the recovery times.  There are less
files to skip over!

The stats you show don't point to any obvious conclusion that I can
see.  Is it repeatable?  While it is running, look for a bottleneck
somewhere, that is all I can suggest.

Regards,
Simon
----------
| From: alex.fagioli /  mime, , , alex.fagioli AT TECTRADE.CO DOT UK
| From: alex.fagioli /  mime, , , alex.fagioli AT TECTRADE.CO DOT UK
| To: ADSM-L /  mime, , , ADSM-L AT VM.MARIST DOT EDU
| Subject: Re: selective vs incremental backup processing times
| Date: Tuesday, 21 December, 1999 4:27PM
|
| What happens when you run an incremental backup
| forever?............................
|
| Once again it should be emphasized that there is no need whatsover to run
| selective backups.
|
|
| Alex Fagioli
|
|
| Data Management Consultant
|
|
|
|
|
|
| Barbara Andrews <bandrews AT ERIE1.WNYRIC DOT ORG> on 20/12/99 20:42:03
|
| Please respond to "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
|
| To:   ADSM-L AT VM.MARIST DOT EDU
| cc:    (bcc: Alex Fagioli/Tectrade Computers Limited)
|
| Subject:  selective vs incremental backup processing times
|
|
|
|
| We are running ADSM server Ver. 3 Rel. 1 Level 2.40 on an OS/390 2.7
| platform.  The ADSM client is Ver. 3 Rel. 1 Level 0.7 running on a Novell
| Netware 4.1.1 Server.  Does anyone have any ideas why it would take much
| longer to run an incremental backup verses a selective backup?  As an
| example, I specified the following object for the backup:
|     vol1:users/ - subdir=yes
| Here is the summary after the SELECTIVE backup:
|
| 12/17/1999 01:05:49 Total number of objects inspected:   33,116
| 12/17/1999 01:05:49 Total number of objects backed up:   33,116
| 12/17/1999 01:05:49 Total number of objects updated:          0
| 12/17/1999 01:05:49 Total number of objects rebound:          0
| 12/17/1999 01:05:49 Total number of objects deleted:          0
| 12/17/1999 01:05:49 Total number of objects failed:           0
| 12/17/1999 01:05:49 Total number of bytes transferred:     1.59 GB
| 12/17/1999 01:05:49 Data transfer time:                1,982.16 sec
| 12/17/1999 01:05:49 Network data transfer rate:          843.95 KB/sec
| 12/17/1999 01:05:49 Aggregate data transfer rate:        500.70 KB/sec
| 12/17/1999 01:05:49 Objects compressed by:                    0%
| 12/17/1999 01:05:49 Elapsed processing time:           00:55:40
|
| Here is the summary after the INCREMENTAL backup the next day:
|
| 12/18/1999 08:13:41 Total number of objects inspected:   33,367
| 12/18/1999 08:13:41 Total number of objects backed up:      227
| 12/18/1999 08:13:41 Total number of objects updated:          0
| 12/18/1999 08:13:41 Total number of objects rebound:          0
| 12/18/1999 08:13:41 Total number of objects deleted:      1,286
| 12/18/1999 08:13:41 Total number of objects failed:           0
| 12/18/1999 08:13:41 Total number of bytes transferred:    55.96 MB
| 12/18/1999 08:13:41 Data transfer time:                  100.83 sec
| 12/18/1999 08:13:41 Network data transfer rate:          568.40 KB/sec
| 12/18/1999 08:13:41 Aggregate data transfer rate:          1.35 KB/sec
| 12/18/1999 08:13:41 Objects compressed by:                    0%
| 12/18/1999 08:13:41 Elapsed processing time:           11:46:30
|
| We also back up another Novell Netware 4.1.1 server incrementally with no
| problems.  It's client code is at a level just prior to the other server's
| ( 0.6).  It will typically run like the following:
|
| Date and Time       Message
| ---------------------------------------------------------------------------
| -------------------------------------------------------------------------
| 12/17/1999 21:00:34 ANE4952I (Session: 267, Node: E011LIB)  Total number of
| objects inspected:   65,642
| 12/17/1999 21:00:34 ANE4954I (Session: 267, Node: E011LIB)  Total number of
| objects backed up:      142
| 12/17/1999 21:00:34 ANE4958I (Session: 267, Node: E011LIB)  Total number of
| objects updated:          0
| 12/17/1999 21:00:34 ANE4960I (Session: 267, Node: E011LIB)  Total number of
| objects rebound:          0
| 12/17/1999 21:00:34 ANE4957I (Session: 267, Node: E011LIB)  Total number of
| objects deleted:          0
| 12/17/1999 21:00:34 ANE4959I (Session: 267, Node: E011LIB)  Total number of
| objects failed:           3
| 12/17/1999 21:00:34 ANE4961I (Session: 267, Node: E011LIB)  Total number of
| bytes transferred:   145.05 MB
| 12/17/1999 21:00:34 ANE4963I (Session: 267, Node: E011LIB)  Data transfer
| time:                  148.33 sec
| 12/17/1999 21:00:34 ANE4966I (Session: 267, Node: E011LIB)  Network data
| transfer rate:        1,001.35 KB/sec
| 12/17/1999 21:00:34 ANE4967I (Session: 267, Node: E011LIB)  Aggregate data
| transfer rate:        236.46 KB/sec
| 12/17/1999 21:00:34 ANE4968I (Session: 267, Node: E011LIB)  Objects
| compressed by:                    0%
| 12/17/1999 21:00:34 ANE4964I (Session: 267, Node: E011LIB)  Elapsed
| processing time:            00:10:28
| 12/17/1999 21:00:35 ANR0403I Session 267 ended for node E011LIB (NetWare).
| 12/17/1999 21:00:37 ANR0406I Session 268 started for node E011LIB (NetWare)
| (Tcp/Ip 168.169.11.20(1341)).
| 12/17/1999 21:00:37 ANR0403I Session 268 ended for node E011LIB (NetWare).
|
| What could be so different about the two servers?
|
| Any insight would be appreciated.  We are only operating ADSM in a test
| mode now, but if we don't get this resolved it may never make it to
| production.  Confidence levels aren't too high around here.
|
<Prev in Thread] Current Thread [Next in Thread>