Results 1 to 5 of 5

Thread: VPS Speed and Connectivity

Hybrid View

Previous Post Previous Post   Next Post Next Post
  1. #1
    Junior Member Newbie
    Join Date
    Jun 2007
    Posts
    2

    Angry VPS Speed and Connectivity

    Is there anything that can be done for more speed and better connection? Sometimes my sites are painfully slow. Although this is a problem we are having with our VPS, it is a connection load issue for all of Spry's hosting.

    Our site along with our customer sites are in a VPS so we fell the pinch on both ends. We not being up and our clients calling wondering what's going on with their sites as well! I can not have this happening as they will eventually become EX-CLIENTS. How can I stay in business, much less a Spry customer, if this happens?

    I've talked and support ticketed this thing to death and it all boils down to Spry's internet provider being overloaded. They have more than one provider but they have to MANUALLY switched over to it (** Corrected. Please see below. -Phil **). Can this not be automated and can't this switching happen as needed instead of AFTER all the Spry customers go down?

    When the connection drops, the servers do not. I know this because my server registers over 30+ days of "uptime". SO, when the connection drops out does this counts against the 100% (99.7 and lower or whatever it actually is) uptime guarantee? Although I've asked twice, I've yet to get a straight answer.

    I really like the service, servers, VPS, the level of access, and most of the things that go along with Spry hosting. If you could just PLEASE get the connection dropping handled you would be perfect and make myself along with all the other people who enjoy your service very happy. Don't you want us happy?

    Please respond. I know a lot of us would like to know.
    Phil
    Last edited by ParaPhil; 06-12-2007 at 01:13 PM.

  2. #2
    Member Newbie
    Join Date
    Oct 2006
    Posts
    46

    Default

    Phil,

    One of our upstream providers did have a major outage on June 4, and there were impacts for many customers related to that outage. Unfortunately, because the outage was off of our network, our uptime guarantee does not apply.

    I think you may have been given the wrong impression about the nature of the manual work that our network engineers did last week. We do have our systems setup to failover automatically, but in the case that a connection to a provider is highly latent or intermittent, that failover may not happen or it may be ineffective on it's own. The failure last week was such an intermittent type of failure upstream, so our network engineers did additional manual configuration in an attempt to limit the impact of the outage for as many remote locations as possible.

    If you've still been experiencings slowness and intermittent connections, we'll need you to submit a traceroute with your ticket. However, I believe that any performance problems you've experienced in the last week are not likely network related. Rather, I have discovered another VPS account on your hardware node that has been consuming an immense amount of the CPU for their VPS.

    Because our attempts to contact that other user and resolve the resource abuse have been unsuccessful, we have recently disabled that account. Please either follow up here or submit a trouble ticket if you continue to experience problems connecting.

  3. #3
    Junior Member Newbie
    Join Date
    Jun 2007
    Posts
    2

    Default

    Toddat,
    Thank you for getting back to us and further explaining the interruption and failover process. Our frustration about the whole June 4th incident was amplified by not knowing, and not being able to get an answer on what exactly was happening. I checked the site and the forums but didn't see anything about the troubles. I'm sure a lot of people were scrambling trying to get the servers back online but a quick line to ease everyone's minds could have been helpful. Just a thought for next time.

    Also, if you disabled the account using all the CPU resources a few hours ago, I can already tell the difference.

    I just checked the resources and see the following for Memory Usage:

    Total: 11.71 GB
    Used: 11.62 GB
    Free: 95.55 MB
    Shared: 0 B
    Buffer: 253.10 MB
    Cached: 3.80 GB
    Usage: 99.2%

    From the Spry site, our Plesk VPS 300 guarantees 512 MB burstable to 1024 MB. I know we're using some of the resources but shouldn't the available be hovering just under the 512 MB mark or does our OS account for some of it's usage?

    Thanks again for getting back to me!
    Phil

  4. #4
    Member Newbie
    Join Date
    Oct 2006
    Posts
    46

    Default

    The info you see when running 'free' in the VPS is not specific for your own VPS. That will show the memory usage of the whole server.

    The particular usage and values for your VPS are found in a special file on your server called /proc/user_beancounters. If you search these forums for 'user_beancounters' or 'privvmpages', you should be able to find plenty of info about memory usage inside of a VPS.

    Additionally, one of our VPSLink customers has written a handy utility to monitor the status of his account in a web-browser. This is for a slightly different underlying technology, but I think the memory info should be displayed properly for a Spry VPS as well.

    Check out labradordata.ca/vpsinfo/ for instructions and a download.

    It was just a few hours back that we disabled the other account on your hardware node. I'm glad to hear that it's helped to improve your performance already.
    Last edited by ToddatSpry; 06-12-2007 at 02:33 PM.

  5. #5
    Junior Member Newbie
    Join Date
    Jun 2007
    Posts
    10

    Default

    Phil

    we have been complaining to Spry tech support about slow web page speeds for months.

    Recently we set up a copy of our website on ProVPS to see if the slow page speeds we were experiencing were due to our code as Spry repeatedly advised or with Spry's service.

    In tests this morning we found that our web pages on the Spry VPS took 22 seconds on average to download. In the same tests at the same time we found that our web pages on a ProVps 384 server took 9 seconds on average to download. This is a 60% performance improvement on a server that is lower spec on paper than our Spry plan.

    As a result we have decided to cancel our hosting service with Spry and move our site to ProVPS.

    So all you need to do to improve your web page speed 60% overnight is move to a better web hosting service.

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •