Post reply

Memory Available Yesterday 11gb today 1gb

  • March 31, 2011 at 8:10 am

    #234857

    Hi,

    So im trying to monitor a cluster that had 3 instances on it.

    I started running a perfmon of the server (Not instance) yesterday. I stopped it today as i found some better things to monitor instead of just Memory: Pages/Sec and Memory Available bytes.

    So before i started to monitor per instance instead of all of them. I toke a look at the csv of my running yesterday ( I had counters for: \Memory\Pages/sec and Memory\Available Bytes)

    So from 30:03:2011 11:00 to 31:03:2011 16:00

    At the start i had: 12166852608 which works out at roughly 11gb

    At the end (20 mins ago or so) it has 1138442240 (about 1gb)

    Pages avg 0 (Which i belive is good and how it should be!)

    I am about to start a new monitor with:

    SQL Server: Memory Manager: Total Server Memory

    SQL Server: Memory Manager: Target Server Memory

    SQL Server Buffer Mgr: Lazy Writes/Sec

    Memory Object: Available Bytes

    Memory Object: Pages/Sec

    I also started a Profiler trace covering:

    Sort Warnings,

    Server Memory change

    Execution Warnings

    So far all ive seen is Sort Warnings (Which i guess is normal?)

    I have found that when someone in the office runs SQLIOSim it wreaks the server memory and seems to keep 10gb or so untill the user fully logs out even though he says the test takes 5mins.. I checked and SQLIOSim was in the Task Manager, But after logging him out im still at 32gb in use on task manager.

    Any suggestions? Currently just reading anything i can find on memory in sql 2008 r2 or below and seeing it can help. but i dont really want to make any changes unless i know its going to help.. (oil on the fire type of thing)

    Thanks for any input.

    S

    Server config is:

    Each instance has min max memory set

    instance 1: 6gb min 12 gb max

    instance 2: 5gb min 10 gb max

    instance 3: 4gb min 08gb max

    Theres 32gb in the server.

    As far as i know. No changes have taken place on the server

  • March 31, 2011 at 11:55 am

    #1306139

    Update,

    I rolled over the cluster to the other node and restarted that machine.

    Memory is now running back at the 10gb mark.

    the perfmon results show that the memory went from 11gb free to 1gb free from 2-5 in the morning..

    Is this just a random act of "god" or something that people see happening a lot.

    Thanks

    S

Viewing 2 posts - 1 through 1 (of 1 total)

You must be logged in to reply to this topic. Login to reply