Rapid Recovery

Poor Backup Performance with AppAssure

Has anyone come up with performance tweaks they can share in AppAssure ? 

I frequently have many jobs run at sub 100K transfer speeds and take ten's of hours to finish. During these times the server is utilizing less than 10% of the processors load, the network shows no more than 5% utilized. I am concerned there is something more going on. I talked to support and they have had me make tweaks, but their last response was why don't you try upgrading to 15K SAS drives and see if this helps.  Not really a solution i want to try without hard numbers to back it up.

I have tested with ATTO and Iometer bench tests and can get line speed transferring data. In addition I went as far as to test Unitrends backup software and i can sustain higher 50MB + transfer speeds for backups. Its not making much sense and though i would check with the group and see if anyone else is experiencing this as well. 

  • Welcome to the world of trouble-shooting with Appassure. The Support tech cant give you any recommendations to find the bottleneck with appassure because there are no logs/ tools that I have ever heard of to help him or you.

    You have to use perfmon (or something like it) to try and get a handle on the bottle neck. And unless the bottle neck is painful obvious (it hardly ever is) then its a huge pain. Even if you find Disk1 IO is off the charts, you wont really know why or what exactly is causing it. That being said, the first things I would look at are memory shortages, keep in mind its normal for a machine to show free memory when in fact this memory is reserved for the OS and NOT available for applications. I find the taskmanager columns pagefaults and page fault delta help a ton to quickly see if applications are memory starved

    I am not familiar with Unitrends so I don't know how they backup but you are probably not comparing apples to apples when you look just at transfer speed. Of course at the end of the day if one product works in your backup window and the other does not. That's apples to apples

  • Thanks, its unfortunate we have to deal with these issues. I m actually considering moving back to MS DPM given our speeds to backup and restore were never an issue. I really just want a backup product that backs up fast and restores fast...i would even give up things like de-duplication and encryption to have this.

  • I love Appassure. If it works, its awesome. The different client restore options are great (Core protection is bad) The problem is, as soon as there is a problem you are pretty much on your own.

    Its a real love hate relationship. Once day I am the hero because I spin up a Hyper-V export 10 seconds after a server failure, But the next week I have every executive in the company sending me emails asking me why is taking 2 weeks to trouble-shoot a simple issue and they want status updates every hour.

    Its often the little things that drive you nuts. Like working on an issue for days and getting no where, your google-fu brings up nothing. Spend more time writing up everything you tried, everything you found and submitting a case. Waiting days for someone to actually look at the case and then they tell you its a known issue but they don't ever bother to take 30 seconds to document it for their customers.

  • dan.lebaron said:

    I frequently have many jobs run at sub 100K transfer speeds and take ten's of hours to finish. During these times the server is utilizing less than 10% of the processors load, the network shows no more than 5% utilized.

    Have you found that this is primarily on the larger servers and happens for both Full (Base) / Incremental backups?

    I find for some reason, the larger servers take longer to backup than the smaller ones.

  • I agree that tuning AppAssure performance is sometimes more and art than a science, but have a look at the following article : documents.software.dell.com/DOC161797

    And more specific : Outstanding Reads per Stream Specifies how many queued read operations will be stored on the back end. This setting helps to control the queuing of agents.

    NOTE: Dell recommends setting this value to 24.  

    I remember having a similar issue where base images were fast, but incrementals were slow. changing this particular setting made a huge difference in this specific case.

    There are some other suggestions in the above document that you can try as well, but as always document your changes