Networker

Re: [Networker] Slow backups

2003-06-12 00:29:09
Subject: Re: [Networker] Slow backups
From: Bokkelkamp Ernst <ernst.bokkelkamp AT SIEMENS DOT COM>
To: NETWORKER AT LISTMAIL.TEMPLE DOT EDU
Date: Thu, 12 Jun 2003 06:25:49 +0200
Hallo Mark,

thanks for mentioning the obvious, I had already forgotten it.

Observing the backup process is something few do, but is one of the most
valuable tools because you can see the overall picture that is not visible
in the log files.

Bye
Ernie

-----Original Message-----
From: Mark Bradshaw (BTOpenWorld) [mailto:notthehoople AT BTOPENWORLD DOT COM]
Sent: Thursday, June 12, 2003 12:36 AM
To: NETWORKER AT LISTMAIL.TEMPLE DOT EDU
Subject: Re: [Networker] Slow backups


Hi Marc,

How are you calculating the throughput rates? I ask as I've just looked into
a suspected slow SAN backup issue which turned out to be a lack of tape
drives - the client in question was taking a long time to get hold of a tape
drive using DDS due to other activity on the backup server which lengthened
the time the backup was running and hence made it look slow. However,
watching the job it showed that once a drive was grabbed then the backup was
perfectly fast!

If you don't have drive issues then things I would look at would be:

- type and configuration settings on E6500 and E10k HBAs
- st.conf settings (if applicable...) on each server. Are they the same?
- SAN switch settings
- other SAN traffic on the same HBAs?
- other processing going on at the same time (batch jobs???)

Also try watching the backups when they first start - you have a good chance
then of seeing both servers backing up database files containing data
(instead of, as others have mentioned, sparse files on one or other server
later in the backup).

Good luck!

Mark

> I have 1 backup that gets a rate of 28GB/hr  and another that is 43GB/hr.
>
> Both are Solaris 6 Servers running Networker 6.1.
> Networker Server is Solaris 8 Networker 6.1.
>
> Both Clients are running Informix and both clients are storage nodes that
> share DDS tape drives across the SAN.
>
> One client is a SUN E6500 (slower) the other is a SUN E10K (faster).
>
> Both have parallelism of 2.
>
> What can I check to determine why one is so much faster than the other?
>
> Thanks,
> Marc Levitan
> Storage Management Specialist
> PFPC Global Funds Services
>
> --
> Note: To sign off this list, send a "signoff networker" command via email
> to listserv AT listmail.temple DOT edu or visit the list's Web site at
> http://listmail.temple.edu/archives/networker.html where you can
> also view and post messages to the list.
> =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=

--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=

--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=

<Prev in Thread] Current Thread [Next in Thread>