cancel
Showing results for 
Search instead for 
Did you mean: 

Muliplexing with NetBackup 6.5.5

kproehl
Level 5

We currently have an environment where we backup directly to lto 4 tapes.  We have a 6 drive library and current;y have multiplexing set to 8 for each drive.  We have had an issue lately where many of our jobs get queued up or error out with 196 errors.  In all cases the error says waiting for tape drives to become available.  I want to increase the maximum number of streams from 8 to 12 for all the tape drives to help fix some of these issues.

1.  Will doing this help fix the the queuing of jobs?

2.  What are the down sides to doing this?  Is there any reason why we should not increase maximum streams from 8 to 12?

Please let me know of you opinion on this.  I want to make this change but I am unsure of the negatives that would occur by doing this.

Thanks.

1 ACCEPTED SOLUTION

Accepted Solutions

Ed_Wilts
Level 6
It really, really depends on where your problems are.  An LTO-4 drive can take data at a great rate but you need the horsepower to be able to feed the drives at that rate.

An LTO-4 has a maximum native transfer rate of 120MB/sec.  That's UNCOMPRESSED data.  Assuming you are getting 2:1 compression, you can fire 240MB/sec at it.  That means you're going to have to dedicate a 4Gbps HBA port for EACH drive.   It also means that your media servers are going to have to be capable of sending them that much data - and I hope you have multiple media servers since you'll need to be able to send out 1.4 GB/sec.

And then you have to be able to suck that much data in from your clients at that rate and that is not going to happen without 10Gbps interfaces or a lot of 1Gbps interfaces.  Assuming you get 70MB/sec out of a single GigE connection, you'll need over 20 network interfaces - (6*240)/70 - to keep the drives busy.

Increasing multiplex rates or adding additional streams only makes sense if you can keep up.

The bottom line is you need to figure out where in the data path the bottlenecks are and then start to address those bottlenecks. 

View solution in original post

4 REPLIES 4

Will_Restore
Level 6

To maximize drive use, multiplex the slower clients that produce small backups.

The higher-performance clients that produce long backups are likely to use drives

fully and not benefit from multiplexing


 

 Allow multiple data streams attribute
 

 

The Allow multiple data streams attribute specifies that NetBackup can divide

automatic backups for each client into multiple jobs. The directives, scripts, or

templates in the backup selection list specify whether each job can back up only

a part of the backup selection list. Since the jobs are in separate data streams,

they can occur concurrently.

The directives, scripts, or templates in the backup selection list determine the

number of streams (backup jobs) that start for each client. The list also determines

how the backup selection list is divided into separate streams.

The following settings determine the number of streams that can run concurrently:

Number of available storage units

Multiplexing settings

Maximum jobs parameters
 

Ed_Wilts
Level 6
It really, really depends on where your problems are.  An LTO-4 drive can take data at a great rate but you need the horsepower to be able to feed the drives at that rate.

An LTO-4 has a maximum native transfer rate of 120MB/sec.  That's UNCOMPRESSED data.  Assuming you are getting 2:1 compression, you can fire 240MB/sec at it.  That means you're going to have to dedicate a 4Gbps HBA port for EACH drive.   It also means that your media servers are going to have to be capable of sending them that much data - and I hope you have multiple media servers since you'll need to be able to send out 1.4 GB/sec.

And then you have to be able to suck that much data in from your clients at that rate and that is not going to happen without 10Gbps interfaces or a lot of 1Gbps interfaces.  Assuming you get 70MB/sec out of a single GigE connection, you'll need over 20 network interfaces - (6*240)/70 - to keep the drives busy.

Increasing multiplex rates or adding additional streams only makes sense if you can keep up.

The bottom line is you need to figure out where in the data path the bottlenecks are and then start to address those bottlenecks. 

richs24
Level 4
also, make sure your media multiplexing value within the policy schedule matches the value you set for your drives

J_H_Is_gone
Level 6

Have you looked while backups are running.

If you have 4 tape drives and mpx set to 8 - then while backups are running you should have 32 jobs running.
So first check drives * mpx setting = active jobs.

If you do not have that many active jobs try to find out why.

Policy could limit might be smaller then drive mpx setting.
or the global setting could be smaller.