Server slows down to a crawl, but works fine after a restart for a while…

Home Forums Server Operating Systems Windows Server 2012 / 2012 R2 Server slows down to a crawl, but works fine after a restart for a while…

This topic contains 9 replies, has 6 voices, and was last updated by  nmoneill 6 months, 2 weeks ago.

Viewing 10 posts - 1 through 10 (of 10 total)
  • Author
    Posts
  • #167534

    Long story short, we have a machine with Win Server 2012 that becomes unusably slow after a period of time. Usually about a week. Meaning you can’t access shared drives or it takes forever, you can’t RDP or teamviewer in, and when I go to the actual server, its not usable either. It will take 10 minutes for the screen to come up for me to enter the user password. Then when I get in, folders and programs say “not responding” often and are basically unusable. Restarting it fixes it and makes it completely normal again for a while.

    It has a RAID5 and the HP RAID software says the drives are good. I did a memory test (memtest64) from within windows for 1 hour and no issues. I ran sfc /scannow and everything is fine with that too. Not sure what’s causing the problem and if its hardware or software related. Any ideas?

    Xeon E5-2620 v4 @ 2.1GHz, 16 gigs ram, 3x 600GB hard drives in a RAID5. Windows Server 2012, symantec endpoint protection, and malwarebytes premium are on it. We basically have a medical program that doesn’t use much resources that run off the server and some basic file sharing with excel, word and pdf files mostly, etc… Also checked logs and don’t see anything weird.


    Ossian
    Moderator
    #191989

    Any application running on it which might have a memory leak? (I am thinking particularly of IIS web applications). I would look closely at the “medical program”

    How about a scheduled overnight reboot?


    sooner555
    Member
    #392018

    Don’t think so, but I will keep an eye on raising ram usage. The medical program hasn’t given us issues in the past and the company has been out of business for years so there haven’t been any updates.

    Is it normal for Domain Name System (DNS) Server to use up almost 600MB of RAM?

    Ossian;n516949 wrote:
    Any application running on it which might have a memory leak? (I am thinking particularly of IIS web applications). I would look closely at the “medical program”

    How about a scheduled overnight reboot?


    Ossian
    Moderator
    #191992

    600Mb for DNS sounds reasonable depending on the size of your network and the queries it is supporting. Is the server also a DC?

    I noticed “RAID 5” in your original post. Probably not an issue, but worth checking for failed / rebuilding RAID drives as that can really slow things down.

    How about setting alerts (via perfmon) on high RAM usage and then look at individual processes consuming a lot.


    sooner555
    Member
    #392019

    I’m not sure if its the domain controller. Unfortunately I know very little about servers. But to try to answer your question (not sure if it is actually answering it), we do have a domain when I look in “system” under control panel. Each computer has local accounts and the accounts are not controlled by the server.

    I looked in the HP RAID software and it said the status of the RAID and the drives is “ok”

    I also setup recording of stats to a log for a week. The alerts thing looked complicated.

    I will update if anything comes up. Thanks!

    Update: I was just running memtest64 and it kicked me off RDP and it won’t connect via teamviewer. Teamviewer was getting stuck on connecting, but now it says its not even running on the server. RDP closes after “estimating connection quality” for a long time. These are the same symptoms (not being able to connect to the server via rdp and teamviewer) as when the server slows down to a crawl. Same thing happened last night as well, but the server was running fine they told me when they got into the office.

    Not sure if its from maxing out the RAM or something else. The weird thing is yesterday while on site I ran the same memory test for 1 hour and I had no problem and it completed successfully all while being connected with RDP the whole time.

    Not sure if this info helps or makes things more confusing

    Ossian;n516954 wrote:
    600Mb for DNS sounds reasonable depending on the size of your network and the queries it is supporting. Is the server also a DC?

    I noticed “RAID 5” in your original post. Probably not an issue, but worth checking for failed / rebuilding RAID drives as that can really slow things down.

    How about setting alerts (via perfmon) on high RAM usage and then look at individual processes consuming a lot.

    Blood
    Blood
    Moderator
    #337383

    It may be worth opening Task Manager, clicking the Details tab, then sorting it by the Memory column or Commit size (see link) and leaving it on (don’t minimise it and don’t log off, just lock it). When you manage to log on to the server when it is slow you should see the offending process at the top of the list.
    Have you seen this: http://social.technet.microsoft.com/Forums/en-US/itprovistanetworking/thread/c3fc9f5d-c073-4a9f-bb3d-b7bb8f893f78/

    #391970

    I’ve come across this periodically at sites that I’ve managed. All of the things suggested by others should be looked into, along with the following:

    1. Definitely keep task manager open as previous suggested so that you can see what’s eating up the memory of CPU. Many times it’ll be IIS or a Windws svc eating up the mem or CPU.

    2. I would remove the Malwarebytes Premium from it as the Endpoint Protection should be enough for it. Both items are probably running realtime scans in the back and fighting for resources all the time.

    3. Do you have any other servers you can offload some roles to or maybe virtualize this server to another machine to see if it’s a hardware constraint?

    4. When the RDP or Teamview disconnect for long periods are other functions of the server still running properly… DNS, the medical application, file sharing… or do they all spaz out and stop working? That could indicate network connectivity issues.

    5. This question will make me sound like a HP or Dell rep… have you updated all of the hardware drivers… BIOS, RAID controllers, network drivers, chipsets, etc. As silly as it sounds and against it when they tell me to do it, I end up with egg on my face as it has resolved issues like this after updating them per their suggestions.

    6. Is the server hardware under warranty? If so, reach out to HP and have them run their diagnostic tools for RAID and other stuff and they will get a log dump and sift through it. Let them do this for you as their is that particular engineers specialty.

    7. Look into getting an RMM monitoring agent on the server like from Atera or Continuum so they will detect anomalies in the performance of the server and alert you to specific issues that might be the ultimate issue.

    Until then, do as the other’s suggest and set a scheduled task for a reboot early morning before the work day starts.

    I hope the questions help and lead you in the right direction.


    sooner555
    Member
    #392020

    I will leave it open, good idea! I haven’t gotten the error code in the article.

    Blood;n516961 wrote:
    It may be worth opening Task Manager, clicking the Details tab, then sorting it by the Memory column or Commit size (see link) and leaving it on (don’t minimise it and don’t log off, just lock it). When you manage to log on to the server when it is slow you should see the offending process at the top of the list.
    Have you seen this: http://social.technet.microsoft.com/Forums/en-US/itprovistanetworking/thread/c3fc9f5d-c073-4a9f-bb3d-b7bb8f893f78/


    sooner555
    Member
    #392021

    1. I’ll leave it open and check
    2. Ok
    3. No unfortunately. We use the server for very little so I don’t think its being maxed out. The problem occurs sometimes even when no one is using the server at all and no one is in the office.
    4. No, they all stop working or become extremely slow. I thought it may be a network issue as well, but even when I go to the server itself, its extremely slow and not usable.
    5. No, does that ever actually fix things when things have been working okay forever? I thought about updating the firmware on the drives, raid controller, and bios, but I’ve also had firmware updates kill hardware in the past so I’m really afraid to. It would really destroy our operation.
    6.I doubt it, its about 2-3 years old. Not sure if server warranties are longer than workstations, but I can take a look.
    7. I’ll look into that.

    I’ll do the reboot thing too, good idea. Thanks!

    Simeon R Naranjit;n516964 wrote:
    I’ve come across this periodically at sites that I’ve managed. All of the things suggested by others should be looked into, along with the following:

    1. Definitely keep task manager open as previous suggested so that you can see what’s eating up the memory of CPU. Many times it’ll be IIS or a Windws svc eating up the mem or CPU.

    2. I would remove the Malwarebytes Premium from it as the Endpoint Protection should be enough for it. Both items are probably running realtime scans in the back and fighting for resources all the time.

    3. Do you have any other servers you can offload some roles to or maybe virtualize this server to another machine to see if it’s a hardware constraint?

    4. When the RDP or Teamview disconnect for long periods are other functions of the server still running properly… DNS, the medical application, file sharing… or do they all spaz out and stop working? That could indicate network connectivity issues.

    5. This question will make me sound like a HP or Dell rep… have you updated all of the hardware drivers… BIOS, RAID controllers, network drivers, chipsets, etc. As silly as it sounds and against it when they tell me to do it, I end up with egg on my face as it has resolved issues like this after updating them per their suggestions.

    6. Is the server hardware under warranty? If so, reach out to HP and have them run their diagnostic tools for RAID and other stuff and they will get a log dump and sift through it. Let them do this for you as their is that particular engineers specialty.

    7. Look into getting an RMM monitoring agent on the server like from Atera or Continuum so they will detect anomalies in the performance of the server and alert you to specific issues that might be the ultimate issue.

    Until then, do as the other’s suggest and set a scheduled task for a reboot early morning before the work day starts.

    I hope the questions help and lead you in the right direction.


    nmoneill
    Member
    #392042

    hello,
    Terminal Services ManagerApplication logSystem logTask Manager and Resource MonitorHP System Homepage or iLO for hardware logRAID controller logs for drive issues
    What software is on that server, and is it getting updates? during these periods? Have users installed anything of their own? Task Manager should show you if any particular processes are using a lot of resources (CPU, RAM, LAN). Check for “”extra”” folders that might indicate the server was hacked and now does P2P file sharing for the Internet! Do you have anti-malware app on the server? Ask 1 user at a time to stay off the server for a week and see if things are better, and investigate anyone causing a lag.
    regards
    patrisia

Viewing 10 posts - 1 through 10 (of 10 total)

You must be logged in to reply to this topic.