Backup exec 12.5 performance issue

Are you backing up local files, or are these files coming across the network? If you are backing up across the network are you using the BackupExec Client?

Are you using the symantec provided tape driver for your backup drive. Even though manufacturers have their own drivers symantec reccomends using their drivers for the tape drive.

I am using nfs storage some of the bakups are local files but some other are over the internal lan network.
I use policy and jobs start automatically

They shold all start at the same time but only one job at time actually becmes active

Of course I also installed the remote agent.  

No isn't network issue all cisco switchs and main routers are configured to 1GBps full duplex.

We double check that. I read something about backup exec and multistream.

Is there any configuratio to allow multistream?

Thanks

I know it can be done but I don't recall doing anything special to force them to run at the same time. Do you only have one B2D folder setup for all of the jobs? I created a separate B2C folder for each job and they run at the same time without issues.

So let's summarize.
* You are doing a Backup to Disk backup to a B2D device that has it's path on an NFS device (is this really NFS or do you mean NAS)? Please specify the device you're backing up to.
* If you run multiple backups they are queued. You ask about multi-stream. Multi-streaming to B2D devices can be done. You need to go into the properties of the B2D device and you'll find an option where you define how many concurrent backups you want to allow to this B2D device.
* You mention differential backups. Differential backups should not be used as a benchmark for performance. Do you get the same poor speed on Full Backups.

Although you seem convinced it's not an network issue, I would suggest you double-check. Every time I've seen this poor of a performance it was a network problem. It's not sufficient that the switches are configured at 1Gbps. You need to check that you've matched that on the NIC as well.
Either set both NIC and switch port to autonegotiate, or fix both NIC and switch to 1000/full duplex. Do not mix autonegotiate settings with fixed speed/duplex settings.

Please answer the above questions and we'll try to sort this out.

I can answer all your questions:

We are backing up using backup to disk folder that has a path on an nfs device. It is not a nas, well.... the function is similar to a nas. Basically the device is a normal server with one 1gb of memory dual core processor xeon and there is a custome nfs OS installed. The os is linux based is from a company called  NASLite-2 v2.10 12-2008
The nas lite is a simple OS for nas system so the device computer in a way has been transformed in a nas server. The network has been checked and all device switchs are configured to autonegotiate.

There is not problems on the network also I am not a genious but networks is my favorites part of IT issue. I am thinking that I configured the jobs wrongly because each job type let's say a full backup job is backedup to his specific path folder called full, then there is a differential backup that is backedup to another path folder called differncial. and so on for all servers that needs to be backed up  so there is many different jobs configured to differents b2d folder. Basically each backup to disk folder has one job configured to a specific path.
Is this the correct way? Probably not, is only couple weeks I am working on storage device

Do you get the same performance on full backup jobs as on differential backup jobs?

Since you say you're running NFS, how did you get the NFS protocol loaded on Windows on the Backup Exec server? Did you install "Services for UNIX" or do you use something else?
I see that NASLite-2 supports SMB/CIFS. Why not use that instead of NFS? It sure would be faster on Windows.

When you copy a large file from the Backup Exec server to the NAS (not using Backup Exec), how fast is the transfer speed?

I only now realize that your first post said that other backups are "superceded". This is probably related to template rules you added to the policy. There is a template rules that says "If start times conflict template A supercedes template B". Probably you have such rules defined in the policy. If you don't want this, just remove the rule.

Maybe you right I think this is the reason why.

I indeed use cifs nas lite automatically create the share called disk-0 by default
Inside that shre thre is all backup folder where all backups are running

I also like to specify that the backup exec is installed on windows server 2003.

Full and diff backup they are both very bad on performance.

The good things is that now I can run more then one job at the same time.

The only issue is the performance it's really bad.
Can somebody has a typical step by step guide line that shows how to configure jobs properly?

Thanks

How fast is the speed when you copy a large file from the Windows server to the NAS?

I am getting 6mb per minutes. This is crap. I am running backup now the issue is the same. Performance sucks and our network admin cannot find any issue on the network.

Try isolating the network issue.
Copy files from another server or client to the NAS. Is the problem identical? That would point to the NAS.
Copy files from the windows server to the client. Is this faster? If it['s fast at least you're sure it's not a problem with the Windows server.
Can you enable an FTP daemon on the NAS and use the FTP protocol instead of SMB. That may indicate if the problem is general or only confined to the SMB protocol.

If you narrow down the problem to a certain server, you could always move the cable for that server to another port on the switch.
And again, make sure you're not mixing autoneg and fixed settings. Either use autoneg on NIC and switch or use fixed speed/duplex on switch and NIC. Do not mix.

Hello all,

I am actually replying to refer this message to techzter;

You are right man this was a network issue, the configuration I done to the backp exec it was brilliant but it was not the cause of the bad performance.

I trusted my network admin and I thought he was right.

I went on site my self and I did check the switchs and the media server. Well...there was a mess some device were conf at 10mb some other 100mb fukl duplex and some other at autonegotiate.

What I did is that I put everything on autonegotiate and now the performance is great even fast then what I thought

Thnaks man the issue is solved