Pinned Client

User photo
Name
Royce Sherman
We have about 2600 pinned clients and have had some issues.  I was wondering if anyone else uses pinned clients like this? 
  1. User photo
    Name
    Jeff Irish

    Royce Sherman:

    We have about 2600 pinned clients and have had some issues.  I was wondering if anyone else uses pinned clients like this? 



    Royce,


    Could you be more specific about the situation that you are experiencing? Is it performance related? Is there a better way that you could doe the same operation?


    In any case, it is difficult for me to provide better advice without more information on the scenario.

  2. User photo
    Name
    Royce Sherman

    We have a network that uses VSat, DSL, and Cell connections.  We have over 1000 different locations spread across the Southeastern U.S. that we provide support. 


    I was wondering if anyone else has had issues using the pinned clients? Performance of the software has been an issue with us.  Because we have so many pinned clients the rep console is slow as a Team Member.  Anything about the Team Member level and your computer crashes.  I have a dual Xeon desktop with 4Gb of memory that crashed in less than 15 minutes.  We found that as long as the pinned client list is collapsed, your computer can run but as soon as you expand the list or search the Rep Console starts eating up CPU time.  It usually bounce between 30-70% usage according to process explorer.

  3. User photo
    Name
    Jeff Irish

    Royce Sherman:

    We have a network that uses VSat, DSL, and Cell connections.  We have over 1000 different locations spread across the Southeastern U.S. that we provide support. 


    I was wondering if anyone else has had issues using the pinned clients? Performance of the software has been an issue with us.  Because we have so many pinned clients the rep console is slow as a Team Member.  Anything about the Team Member level and your computer crashes.  I have a dual Xeon desktop with 4Gb of memory that crashed in less than 15 minutes.  We found that as long as the pinned client list is collapsed, your computer can run but as soon as you expand the list or search the Rep Console starts eating up CPU time.  It usually bounce between 30-70% usage according to process explorer.



    Okay, now I see the issue, or I think that I do.


    Instead of Pinning, install a Jump Client. The advantage here is that the connection is available should you need it, but it is far more configurable on what statistics and information is delivered to the Rep Client. Pinning is going to cause a problem because when the user expands the list, the Rep Client has to display the screen for all of the 2,600 desktops.


    Using the Jump Client, you can determine which Team you want to Pin this to. (If you don’t use Teams, setting them up is easy and you could do it by region or other logical divisions.) Additionally, you can setup what statistics are brought back to the Rep Client and how often. It seems like a win-win. Here’s what the screen looks like. Have a look ans see if this fits your needs. Post more questions as you deem necessary.

  4. User photo
    Name
    Royce Sherman

    Actually, that is what we use to deploy the client out to the sites.  The only statistics we bring back are CPU & Uptime.  The update interval is set to 1 hr.
    Our support center needs access to all of the computers.

  5. User photo
    Name
    Jeff Irish

    Royce Sherman:

    Actually, that is what we use to deploy the client out to the sites.  The only statistics we bring back are CPU & Uptime.  The update interval is set to 1 hr.
    Our support center needs access to all of the computers.



    Hmm. This presents a different issue then. 2,600 Jump Clients is quite a few.


    To have that large of number of Jump Clients you would need to be running a B400 or at least 3 B300’s or B200’s  to handle that load. What is your configuration? If this is on a single B200 or B300 I would be willing to be that the CPU activity spikes when one of your Rep’s expands the listing.


    I know from experience that there is not a way to “balance” a load like this across multiple B300’s, yet. However, you could again break it out by region if possible to allocate your Jump Points.


    I would recommend that you contact Bomgar Support on this, but only if you are running a B400 Enterprise version. They may or may not be in a position to assist you if all 2,600 Jump Points are on a B200 or B300

  6. User photo
    Name
    Royce Sherman

    We have two B300s, but are waiting for a stable code that can handle 2600 jump clients before we look a the B400.  The issue is not on the B300 but on the computer that is running the Rep Console. 

  7. User photo
    Name
    Jeff Irish

    Royce Sherman:

    We have two B300s, but are waiting for a stable code that can handle 2600 jump clients before we look a the B400.  The issue is not on the B300 but on the computer that is running the Rep Console. 

    Oh, so you are saying that it only happens to some users running the Rep Client, but not all? I guess I missed that when reveiwing the scenario.


    I’m really curious about this one now.

  8. User photo
    Name
    Royce Sherman

    No, when all users run the rep console it is doing this.  Most of the users collapse and expand the list as they need to get in to sites.  All of the reps are set as Team Members.  I have a special admin login that I use only to monitor.  It is set to Team Manager.  On my non-admin user profile, if I leave the list expanded for more than 30 minutes, my CPU usage max’s out, with the rep-console leading the way.


    The main reason I started this thread was to see if anyone else use Bomgar in this way.

  9. User photo
    Name
    Jeff Irish

    Royce Sherman:

    No, when all users run the rep console it is doing this.  Most of the users collapse and expand the list as they need to get in to sites.  All of the reps are set as Team Members.  I have a special admin login that I use only to monitor.  It is set to Team Manager.  On my non-admin user profile, if I leave the list expanded for more than 30 minutes, my CPU usage max’s out, with the rep-console leading the way.


    The main reason I started this thread was to see if anyone else use Bomgar in this way.

    We do use it this way, but not for so many locations. Most of the time we use Jumppoints and then connect to servers and workstations via that method. Given what you’ve descibed, I don’t think that would work for your case.



  10. User photo
    Name
    Franky Smith

    I believe the threshold is 2000 jumpclients (or pinned). We make use of a B300 running 4.9.6 and used to have issues with the the clients dissapearing or  going offline. As Bomgar connections are outbound and as we the jumpclients appear to be unstable, we did ask for Bomgar to develop a new connectcion method so that a client can be installed on the remote system that is waiting for an inbound connection from the B300 box instructing it to connect outbound, like the Bomgar button would do.
     

  11. User photo
    Name
    Mark Mitchell

    Bit late to the party here sorry !



    This is a major area of concern for me as my company grows this is the one area that just doesn't appear to be very scaleable.  I notice this with 100 jump clients never mind 2000!



    Something needs doing to improve this - my guess is that the rep client polls the Bomgar for the information and receives back an xml file or similar which is then parsed(could be talking rubbish here but it's how it appears to be working) - it could be this parsing mechanism that's causing the issue especially with high numbers of pinned clients.



    Bomgar really need to look into this - either using a locally stored cache which is refreshed as requested i.e when drilling right down for example or improving the speed at which the Bomgar can serve the information to the clients.



    Just my thoughts !

  12. User photo
    Name
    Darren Ilston

    Mark Mitchell:

    Bit late to the party here sorry !



    This is a major area of concern for me as my company grows this is the one area that just doesn't appear to be very scaleable.  I notice this with 100 jump clients never mind 2000!



    Something needs doing to improve this - my guess is that the rep client polls the Bomgar for the information and receives back an xml file or similar which is then parsed(could be talking rubbish here but it's how it appears to be working) - it could be this parsing mechanism that's causing the issue especially with high numbers of pinned clients.



    Bomgar really need to look into this - either using a locally stored cache which is refreshed as requested i.e when drilling right down for example or improving the speed at which the Bomgar can serve the information to the clients.



    Just my thoughts !

    What connection are you using currently Mark ?



    Our kit in the DC doesn't have these issues, occasionally we get the odd dropped client on a 2 meg SDSL but we put some QOS in place and we "think" that's fixed.

  13. User photo
    Name
    Larry Rathjen

    We have had issues with the jump clients as well.



    The first is that when you try to jump to a client, it just sits there and does nothing. You can RDP into it all day long. It is a hit n miss on different platforms. Bomgar Support says that the console session 0 is not getting freed up.



    The second and more concerning issue is that once you deploy jump clients is that they are now showing as offline. Again, you can RDP into the machine just fine.



    The fix for the console session 0 is supposedly 4-6 weeks out. We have had this product for a little over a month and have deployed around 1K jump clients. Sad to say that we had to write our own internal application to do this. You would think throwing 20K at Bomgar that they would have something already incorporated.


  14. User photo
    Name
    Tony Dahle



    We have roughly 1000 Jump Clients and are not having any issues.



    But, we have different support teams for our techs to use.  And within each team the clients are group by either contract number or location.  This way the tech sees a bunch of locations or contract numbers.  After finding the one they are looking for, they open the smaller sub set to see the actual jump clients.



    You may want to try grouping things as much as possible.  We currently have only 10 jump clients or less in each sub group.  But some of the support teams have 30 groups.



    I don't know if this will help you out or not, but good luck.


Register or log in to leave a reply to this thread.