Random false "offline" status on Windows PCM Dashboard 2.5

[jbohnenstiehl]

This is more annoying than anything. I've been using PC Monitor for a few weeks now and have been pretty impressed with it's usefulness so far. One thing that is nagging me is that randomly throughout the day any of my currently 24 monitored servers may show that it is offline. Obviously I wouldn't be writing if this were the case. Actually as I was typing this they all disappeared from my PCM Dashboard and then just showed back up a few seconds later. I am the Infrastructure Manager here so I would know if we were having any network connectivity problems and certainly if a server were actually offline. No "server offline" tickets are filed by MPCM support emails so I assume this is limited to a Dashboard problem. Furthermore when a server shows as offline on the Dashboard it does not show as offline on my Android MPCM app. Any help with this is appreciated.

Jason

[Paul]

A server is offline event will only fire in the event of network connectivity problems between the Host application and the remote server for more than one minute. The strange thing is that it does not show offline on your mobile application. The information you provided is not enough for a complete understanding of the situation:

• Do you get a server is offline PUSH notification at the time of the problem?
  
• Does that machine happen to perform any other network problems with other services?
  
• Have you tried the web interface of pc monitor to see how is that machine showing up?
  

If this is indeed a Dashboard bug I believe that it shouldn't be limited to one customer, I've been using the Dashboard app too with no problems so far for 4 servers.

[Marius]

Few notes on this before we can start investigating the issue:

- When a server is offline for more than 15 minutes you will get a notification;

- Dashboard reads a state of the computers every 1 minute and if it says offline means the computer was actually considered offline at that time;

- A computer gets the offline state if it has not contacted the server for 20+ seconds;

- Not sure how all your computers just disappeared from the Dashboard (unless the dashboard could not connect)

Please email us at support@pulseway.com and we will enable the debug log on one or two computers to see what is actually going on.

[jbohnenstiehl]

• Do you get a server is offline PUSH notification at the time of the problem? - No
  
• Does that machine happen to perform any other network problems with other services? - No
  
• Have you tried the web interface of pc monitor to see how is that machine showing up? - Yes, same as my Android app, works fine.
  

Marius & Paul, thank you for your quick responses. On Friday I re-installed the Dashboard. I hope to see that my issues are resolved today. I will let you as well as Support know if not. I decided to re-install based off of your responses and the fact that my coworkers do not have the same problems that I have with the Dashboard. Thanks again.

[pedroza23]

I actually monitor 28 server ad i have the same exact problem as jbohnenstiehl.

I seen the problem throught all versions of dashboard since i use pcmonitor. i dont remember exactly but i think it was july 2011.

I see random pc being offline for 1 minute and then comes back online.

One of the 28 servers is a parallels plesk vm in the aruba.it server farm with reduntdant 100mbit connection and i'm pretty sure it's impossibile it loose connectivity so frequently.

However it would be a pleasure if i could set the number of seconds for a server to be set as offline. It would be useful even for really poor and overutilized adsl connections

[Marius]

How often does this happen?

[jbohnenstiehl]

The issue is still occurring for me after I re-installed the software. Sometimes It's one server sometimes it's more. It happens everyday but I can't say that it is at a specific interval. If need be I will record the times. As I was typing this two servers went offline in a site with 9 monitored and also one other server at a different site. I wrote down the time in case it's helpful. I'll continue to do so. Let me know if there is anything else that I can do to assist in troubleshooting. Thanks

[Paul]

At the time of downtime if you have remote access to the computer, can you try connection with telnet to the server to see if there is a connection problem?

terminal ws18.pulseway.com 443

If you get a timeout then the problem is network overflow or worse.

[Marius]

This will happen at least once a day - the application pools are recycled automatically around 5AM GMT. As well, we might deploy a new version of the server as well.

It usually takes 10-20 seconds for all the computers to appear online again and if the Dashboard just did a refresh it will see the computers offline for one minute as the refresh interval for the Dashboard is 1 minute. When this happens press F5 to force a refresh.

[jbohnenstiehl]

This will happen at least once a day - the application pools are recycled automatically around 5AM GMT. As well, we might deploy a new version of the server as well. It usually takes 10-20 seconds for all the computers to appear online again and if the Dashboard just did a refresh it will see the computers offline for one minute as the refresh interval for the Dashboard is 1 minute. When this happens press F5 to force a refresh.

This is happening all day long. Not once a day. Hitting F5 is not an acceptable solution as it only sometimes works. Then I have to pull up a command prompt to ping the "offline" server. It's never offline. However, I understand what you are explaining about the refresh intervals. What do you mean, "we may deploy a new version of the server as well"?

[Marius]

Is it happening for one particular computer? We will need to investigate that and enable debug log. And F5 is not a solution, it's just a way of forcing a data refresh.

[jbohnenstiehl]

Is it happening for one particular computer? We will need to investigate that and enable debug log. And F5 is not a solution, it's just a way of forcing a data refresh.

It is not any specific monitored server. Any of the 26 I have monitored. It seems to be worse today than usual.

How do I enable debug logging?

[Marius]

I have sent you an email with instructions on how to enable the log.

We have been monitoring the Dashboard for the last couple of days and we had no issues.

[jbohnenstiehl]

FYI - I did not feel that the suggested troubleshooting i.e., installing a Beta version of PC monitor onto a server would have helped with my very isolated Dashboard issue.

I again uninstalled the Dashboard, restarted, installed and again restarted and have not been having the same issue with servers showing up as offline when they are not.

I do want to add that I decided I had to do this because three days ago when I would click onto a server to adjust it's setting from the dashboard I got an error message saying that the software was not up to day when in fact it was version 2.6.

Either way I hope the same problem does not resurface but I can say for now that it seems to be resolved. Thanks

[Marius]

Turned out the issue was on the PC Monitor agent so the beta actually fixed the issue.

Dashboard is just a "dumb" client - in 99.9% of cases all the computers state / info comes from the server so if one computer appears offline it's not the Dashboard that has an issue.

[jbohnenstiehl]

I hope you are right. Just now I had 8 servers appear as offline. I'm going to install the beta software on a few servers and then monitor the situation. Do you have any idea when the new version of PC Monitor will be released?

[Marius]

2.6 has been released a few days ago - what is the PC Monitor version on those servers?