pc Updating client data

[Jeroen Aarts]

I have an issue at one of our customers, where clients are not seeing updates in records from each other. Updates from committed records are not pushed to the client, and/or local cache is not refreshed. Even a script with  Commit Record, Refresh Window [Flush cached join results] and a Flush Cach to Disk script steps does not update the data, This is particulary annoying as the application has a realtime planning module. 

 

The issue appears at some intervals, and only restarting the client often resolves the issue for a while, but then again there is a risk that the issue reappears. Restarting the server resolves the issue for a few weeks, but then again it start to appear.

 

The databases are hostned on a partition on an SSD drive on a virtual dedicated Win 2008 R2 machine. It has 8GB RAM and 3 processor cores assigned. There is enough disk space on all partitions. The system and network is managed by the IT department of our customer.

 

Has anyone had similar experiences? Could the SSD drive be an issue?

[David Jondreau]

Is your script calling the Commit, Refresh, etc running on the source client or the target client?

[Jeroen Aarts]

Is your script calling the Commit, Refresh, etc running on the source client or the target client?

 

David, thanks for your reply. I call it on the target client, if you mean the client that is not viewing the most up-to-date data....

[Wim Decorte]

The records are there?  (as in: they don't accidentally enter data in find mode, or in another copy of the file,...)

 

To help rule out variables, I would move the virtual machine to another host and see if the issue persists.

 

What the OS and hardware of the clients and the server (and host machine specs), what FM version?

Anything in the Windows event logs of that virtual machine?

[Jeroen Aarts]

Wim,

 

Yes, the records are there, it is just that clients don't get updates of these records. They are not entering data in Find Mode and they all work in the same file. I stood behind

 

We have checked the Windows event logs but do not find anything of suspect.

 

I also did a Recovery to check for corruption, and there was a corrupted index in another table. But I don't see how file corruption could raise the issue that we are experiencing.

 

Another point of concern seems to me congestion of network traffic, which could lead to FileMaker Server outbound data not being able to reach clients in time. I don't know if this is a plausible scenario?! The network switches do not allow to prioritize network traffic, so we are not able to optimize here.

 

We will move the FMS to a new server- a real machine - in the near future. I hope this resolves the issue but I am not sure as we cannot pinpoint the problem.

[Wim Decorte]

Wim,

 

Another point of concern seems to me congestion of network traffic, which could lead to FileMaker Server outbound data not being able to reach clients in time. I don't know if this is a plausible scenario?! The network switches do not allow to prioritize network traffic, so we are not able to optimize here.

 

 

 

Yes, that is a big concern since FM is a lot more "chatty" than most other things on the network.  If VOIP i used on the same network then that typically has QoS settings that give it priority (to avoid dropped calls and bad call quality) and can push FM to the background and interfere with it.

Same if say there is IP Video monitoring, lots of Netflix/YouTube watching on the network,...

 

Typically though you'd see a good number of disconnects in the FMS event log too if this is the case.