Hi Team,
In an App I have the following SQL that runs as part of the FormCreate.
{sql}
SCRIPT (OUT SessionID INTEGER)
BEGIN
  Set SessionID = CURRENT_SESSIONID();
END
{sql}

In the EDB System Events Log I am seeing messages like the attached..

Is/should the 'CURRENT_SESSIONID()' be the same as reported as 'session ID' in the Description column??

Regards & TIA,
Ian

Ian


Wild guess time from the jpg supplied. The engine can't report the fact that session 6 is no longer there on session 6 so it uses the only session it can be pretty certain exists - session 1 which (again wild guess time) is probably the server.

However, to answer your question - probably unless there's an error like 1107

Roy Lambert

Hi Roy,
I am trying to correlate some Application log entries with the System Events Log.
They are all 1107.
The User didn't use the App for 30 minutes.  Then when he went to use it he got the RemoteReconnect error to start with, handled in the code, then it gave him the 1107.

This of course suggests that at some point in the half hour or so the App & Server lost connection for more than 3 minutes so the server closed the session.  Then the User tried to use the App again and of course it failed.
I initially thought that perhaps  the User had his PC set with auto sleep/standby but this was not the case, so I am still trying to ascertain why the disconnect.

If it is something on the LAN then I would have thought a RemotePing would have triggered a timeout for the App.

Ian

Ian


This is a bit of a where do I start type problem

Possible assumptions:

1.    hardware problem
2.    person problem
3.    there is a problem in your software
4.    there is a problem in the server code

Is this all users (I noticed several session IDs in the jpg) or just one? If just one then assumption 1 or 2 is probably correct, if loads then probably 3 or 4. If 4 then I'd guess Tim would be bombarded with c/s users screaming so most likely 3. However, it could still be hardware.

If the correct assumption is 1 then you need some diagnostic software stuffing on the PC (or PCs) in question - something like wireshark (looooong time since I used it) which should let you spot and network traffic problems. A quick reboot, sleep, hibernate would also cause the connection to be lost, as would (I'm guessing) exiting your application. Other things can be the network card glitching, power spikes, someone tripping over the cable, dodgy connector etc. If its a wireless or mixed wired/wireless then the wireless part can drop out for loads of reasons.

None of this particularly helpful. Can you post the server options you have set? Someone more skilled than I on servers may spot something. The other thing I'd be interested in seeing is the network topology, and a software inventory on any affected PCs. The latter because there is some s/w which will bollix a network connection.

Roy Lambert

On 6/22/2022 3:51 AM, Ian Branch wrote:
> I am trying to correlate some Application log entries with the System Events Log.
> They are all 1107.
> The User didn't use the App for 30 minutes.  Then when he went to use it he got the RemoteReconnect error to start with, handled in the code, then it gave him the 1107.

Would need to know bit more - was it on screen already or only was shown
when user tried to use the app ?


> This of course suggests that at some point in the half hour or so the App & Server lost connection for more than 3 minutes so the server closed the session.  Then the User tried to use the App again and of course it failed.

Assuming you use defaults then session would simply be disconnected at
180 sec mark but still be on the server. Default dead session i think is
another 5 min so in general it would be 8-9 min for session to purge.

> I initially thought that perhaps  the User had his PC set with auto sleep/standby but this was not the case, so I am still trying to ascertain why the disconnect.

Do you believe they are correct - we had a user once who kept closing
laptop lid (which sleeps the laptop) and still telling us laptop was not
in sleep mode.

> If it is something on the LAN then I would have thought a RemotePing would have triggered a timeout for the App.

Yes it should as long as there was not an extended network outage (which
could be from system or network card sleeping).

How are are handling the reconnect - do you prompt user or do silently
in the code ?

If you prompt the user before attempting to reconnect then various
timeouts come into play again since server would wipe the session after
8-9 min anyways. If you do it silently and it generally works then my
cote is for extended network outage again.

Raul

Hi Team,
So, after some interrogation, like pulling hen's teeth, more has been revealed.

Yes, several Users had the issue at approx the same time.  Got the reconnect dialog, clicked the 'Continue' button and all was good on earth..

The User was in the App, went and did something else, came back and the reconnect dialog was on the screen.

Dead session timeout - From the good book "Information If all of the remote sessions accessing the ElevateDB Server are using pinging, then you should set this configuration entry to the minimum value of 10 seconds so that sessions are removed as soon as they stop pinging the server.

The default value is 300 seconds, or 5 minutes."
I have it set at  the default 300 sec.

The User did not have there PC set for auto sleep/standby.

The reconnect dialog is per Tim's suggestion..  Continue, Ignore, Stop.

PC or Lap Top??  Good question, I will ask.

Regards,
Ian

Ian


Just having a look through the newsgroups and it seems as though this problem has been around for some time now, at least there is a one post dating to 2019. I doubt that your client (or even you) have the records to show frequency of the events.  Do you have a feeling that there is any pattern to it, or is there a cluster of events which is followed by peaceful operation then another cluster, or is it a continuous and ongoing problem?

Roy Lambert

Hi Roy,
The Sessions Log does give some indication, as well as the error reports.
They seem to be short, infrequent events.
I am suspecting the Server PC, the Router, or something else on their LAN.

They don't seem overly concerned about it.  Probably used to it.
I will keep monitoring.

Regards,
Ian

Ian

>The Sessions Log does give some indication, as well as the error reports.
>They seem to be short, infrequent events.
>I am suspecting the Server PC, the Router, or something else on their LAN.

Hardware is a prime candidate. If it had just been one PC its pretty much a dead cert.

One thing that crosses my mind is power. A short blip can be enough to knock the network connections out with a PC not noticing. If it was regular it could point to some piece of machinery starting or stopping (had that happen to one company I worked).

Hmmm is it always the same group of PCs or is it random? Any of the PCs had a dual NIC? Is it wireless or wired?


Roy