Video Screencast Help
Search Video Help Close Back
to help
Not able to make it to Vision this year? Get a sampling in the Best of Vision on Demand group.

Symptoms of an Overtaxed Master Server?

Created: 24 May 2010 | 3 comments
rsamora@eprod.com's picture
0 0 Votes
Login to vote

I am trying to determine if my Master Server has finally been overloaded. 

I have two libraries backing up 600 clients: 580 Windows and 20 UNIX: 
Master is Windows Server 2003 x64 SP2 with 24GB of memory. 

SUN STL SL8500
     36 Tape Drives shared with 5 Media servers and 45 SAN Media Servers. 
     Multiplexing set to 1 on 1/3 of the clients and anywhere from 4 to 8 on the rest of the clients.

I recently added an HP Storageworks 9000 that is currently configured with 110 drives but right now only 71 of the drives are assigned to to different media servers.
1 Media Server - 36 drive.
12 SAN Media Servers - 35 drives distributed distributed amongst them in various numbers. 
No shared drives on the virtual library.

Schedule:
5:00 PM - 10 jobs begin
6:00 PM - 55 jobs
7:00 PM - 33 jobs
7:30 PM - 75 jobs
8:00 PM - 60 jobs
9:00 PM - 110 jobs
9:30 PM - 140 jobs
11:00 PM - 150 jobs
12:00 AM - 90 jobs

There are a total of 930 jobs that run from a 5:00 PM to 5:00 AM backup window with the remainder of the jobs scattered in smaller numbers at various times other than the ones shown above.

1/1/2010 through 4/14/2010 - 560 media errors (either status 98 or 84). 
I added the virtual library on 4/15/2010
4/15/2010 to present - 554  media errors (either status 98 or 84).  

That's what makes me think I finally killed my Master Server.

I'm going to open a call with Symantec but I wanted to find out if anyone has a similar load running on a single master server.

Thanks in advance for any input.
Randy

 

Comments

wr's picture
24
May
2010
0 Votes 0
Login to vote

Media errors point to

Media errors point to problems with the Media server(s), not the Master.

good Will backing-up

rsamora@eprod.com's picture
24
May
2010
0 Votes 0
Login to vote

Makes sense.  I still have no

Makes sense.  I still have no idea what to do but maybe I'm homing in on the problem.  Thanks.

rsamora@eprod.com's picture
24
May
2010
0 Votes 0
Login to vote

Update

I just realized that the number of media errors I listed above are for jobs that ended with either a status 84 or 98.  That doesn't take into account the jobs that failed and requeued.  I have the retry set at 4 so the failures were in the 1000's.  Geez!!