Data service issues since v28 upgrade

SOLVED

We're unable to access Sage data on various computers every morning. It seems to affect about half of the office each time, not necessarily the same computers each day. A couple of restarts usually resolves the problem, also tried reinstalling, clearing the .NET files and clear DNS cache, all mixed results. I find it unlikely it's the server because several PCs connect OK at the time, but to have several PCs develop the same problem at the same time seems equally unlikely, but this began to happen immediately following the v28 upgrade last week.

I can see from the company selection screen that the sign-in will fail, the clue is in the displayed version:

If the version is a flat 28.0.0.0 then we know that the Sage 50 app is not happy. First we see this:

Followed by this:

The data service (and control service) is running on the local PC and on the server.

After a restart or two (for this screenshot it was 2 restarts) we see the version correctly reported:

and we now find we can sign-in normally.

I've been following the guidance here:

Troubleshooting data service issues (sage.com)

Can't follow it to the letter as the server is in use (and Sage behaving) so for example I haven't deleted queue.dta yet.

I've also read this:

(+) accounts service issues since V28 upgrade - Sage 50 Accounts and Sage 50cloud Accounts UK General Discussion - Sage 50 Accounts and Sage 50cloud Accounts UK - Sage City Community

and tried the repair option but it's still happening. Any other ideas?

Parents
  • 0

    Hi . Thanks for contacting Sage.

    One potential cause of this issue is if your local or server PC times are not set correctly (i.e., PC clock is out of sync). Can you please verify if the local and server clocks are set correctly and in sync?

    The other Sage City issue you referenced was caused due to usage of mapped drive and was resolved by using a UNC path to the datasets on the server. Not sure if this is the same for you, but worth checking that too please.

    Was it restarting only the client machines that resolved the issue, or did you need to restart the server machines too?

    Hope this helps.

    Thanks.

  • 0 in reply to Ali Al-Amiri

    Hi Ali, yes the time is set correctly and the local clock is in sync with the server clock.

    I've checked for mapped drives and that's not it, we're using UNC paths in the company file and throughout sage.ini. The company selection screen also shows a UNC path.

    It's only restarting the client machines that's needed, never the server, but the client machines might take several restarts before they can reconnect.

    Thanks

    Rob

  • 0 in reply to Rob Daniels

    Thanks for the info Rob.

    Can you please check the time is in sync with the server clock when you see the issue again on one of the clients? I'm just wondering if the client clock is getting out of sync somehow and then a restart is correcting that automatically.

    Have you got any third party Antivirus software installed on the client machine?

    Thanks.

  • 0 in reply to Ali Al-Amiri

    Will do.

    We've had 3 that won't connect this morning (too late for the clock check I'm afraid).

    On the first one I cleared the DNS cache and the .NET folder, then restarted twice before Sage started to connect normally.

    On the second, Sage connected after clearing the DNS cache and the .NET folder.

    I cleared the DNS cache and .NET folder on the third one but still needed 1 restart before Sage connected.

    During this time, 3 other computers were using Sage as normal. I already uninstalled or Antivirus from one of them as a test.

    Thanks

  • 0 in reply to Rob Daniels

    Hi Ali

    This morning we've had 3 computers unable to access Sage data, different 3 to yesterday. All 3 local clocks matched the server perfectly. One of these 3 was the one I uninstalled Antivirus so it seems that's not it either.

    It's beginning to look like clearing the DNS cache might be the quick fix, I restarted the first one before I flushed DNS but it worked OK on it's own for the second and third.

    What would be the permanent fix for this?

    Thanks

  • 0 in reply to Ali Al-Amiri

    Possible workaround

    Since it looks like this might be a DNS issue, I've added the company to the list for a second time but using the IP address instead of the server name. it doesn't fix the cause but so far it allows us to get some work done:

    As you can see here even when the current problem is preventing us from connecting using the server name, the company that uses the IP address is still able to establish the data version and subsequently connects as normal.

  • +1 in reply to Rob Daniels
    verified answer

    Thanks a lot for the extra information Rob, appreciate it.

    Looks like for some reason the client is not resolving machine names to their IP address.

    Can I please ask you to add the following system environment variable to a client PC that has the issue when using machine names in the path and see if that resolves it. We explicitly set that flag on the app so it should be getting automatically picked up, but I think it's worth manually adding it as an environment variable to see if it makes a difference

    Add GRPC_DNS_RESOLVER variable name and set its value to "native" (without the quotes)

  • 0 in reply to Ali Al-Amiri

    Quick Update

    No further issues noted since the environment variable was added to each PC a little over 2 weeks ago. Looks like this is resolved, Thanks Ali

    Rob

  • 0 in reply to Ali Al-Amiri

    Is this added as a User or System environment variable?

Reply Children
  • 0 in reply to Stephen Wade

    Hi Stephen. It would be system environment variables. In practice you shouldn't need to add this variable as the Sage 50 service adds it automatically, but there appears to be instances as in Rob's case where that's not having an effect. Possibly related to specific system environment/setup. I will investigate further to see if I can replicate it, but in the meantime the above workaround can be used. Hope that helps. Thanks.

  • 0 in reply to Ali Al-Amiri

    Interestingly, I have Accounts (v28) and Payroll (v28) installed on my Windows 10 PC and this environment variable did not exist.  So it had not been added during the installation of the data service on my machine.  However, I have not had the above error!

  • 0 in reply to Stephen Wade

    Hi Stephen. Yes when the data service adds the variable automatically you won't see it in the environment variables UI because it is added to the data service process memory only. It acts as a session variable while the process runs rather than a persisted global one.