No announcement yet.

Server loads :(

  • Filter
  • Time
  • Show
Clear All
new posts

  • Server loads :(

    Hello Chaps,

    I run multiple websites on my cloud, and I don't consider any of them to be a major stress on the system - at least, they haven't been until recently. Everything seemed ok until the recent outages on the cluster a couple of weeks ago - now i'm suffering from occassional drops/ time out and very slow page loads.

    I've contacted Live Support numerous times - who tell me that my DB needed optimising one time, then that Apache needed restarting the next.

    Whilst doing a restart solves the problem on a temporary basis, it doesn't resolve this fully.

    My concern is that it is hard to pin down what is causing the issue and whose fault it falls at. I've tried turning off elements to my site for periods that I consider to be more resource hungry - A forum, for example. However, I don't have huge amounts of traffic - not when compared to other major forums that exist. In the back of my mind, I don't feel reassured that the hardware itself is running at an optimal level.

    My question is this: Would EUKHost consider running an analysis service or at least provide a table of what a configuration would be recommended to run? I have upgraded my RAM to 1GB a while back and have just contacted Live Support to begrudgingly upgrade the CPU speed to 1.5Ghz - I'm hoping this makes a difference.. but if it doesn't, I don't know what my next step is?

    Anyone else having issues recently? Anyone have any ideas?

    Thank you

  • #2

    I notice variations in response since the outage , but to be fair it was better for a few days. I am on cluster #3 , and have 2 VM's.
    1 VM is pretty populated, and the other only a few sites .
    With the above VM's , they both stutter about the same , which leads me to believe more hardware / software config on the host side of things, not the guest.

    My CPU is 2 core , 2.7ghz with 2GB ram on each. I do get high I/O , and suspected the feed to the storage several weeks before the outage, plus brought this up with support, but was told everything was fine.

    Over time I have noticed a degradation of the cloud VM's speed. At first when I got them , they were pretty fluid , but now , even when the load is virtually 0 , the cpanel and WHM page speeds are rubbish.
    I also have a dedi with EUK and serveral others dotted around the place with other providers and they are fine most of the time and suffer no lag that I experience with the 2 VM's. Although , apples and pears , the specs are not the same.

    Logging into shell at times it slow , even when load is low and network routing is fine.
    Pings and trace's are fine, I have tried 3 ISP's and they are all the same, plus tried connecting via shell through my other boxes.

    So my conclusion is the VM's are not functioning 100% efficiently.

    I have noticed that immediately after the outage that my speed did increase and the I/O wait was low pretty much of the time, but now it seems like back to how it was.
    I can tell things take longer as my nightly backups reflect what I have experienced , they went shorter to do the backup and now are taking longer again.
    Not quite sure what to do really , certainly not impressed with the latest outage , although I do hand a bit of sympathy as I understand these things do happen. It was said the Titanic was unsinkable ........

    Just my 2p worth...


    • #3
      @ Jester
      I am having a look at your server. Will post further updates shortly.

      MSN :: cristiano @
      Skype :: cristiano.dawson


      • #4
        Hi James.. As per our chat conversation, we found that the CPU and memory were reaching the threshold set. We have performed mysql tweaks and upgraded CPU as requested in the ticket. Gimme a shout if you still face any issues

        MSN :: cristiano @
        Skype :: cristiano.dawson


        • #5
          Thanks Cristiano,

          I feel like i've made real progress towards getting the server running faster. Obviously, boosting up the CPU was going to help - but also deleting old databases and restricting access to certain countries seems to be helping restrict non-useful traffic.

          Thank you for your patience, you are an asset to EUKHost.