10-12-2018 11:20 AM - edited 10-12-2018 11:22 AM
So here is another strange thing its happening to me in the new version (2018) of M.Apps Enterprise. I noticed it while debugging the error I´m getting from the workflows and it´s related with the time shown in the Logs, it appears that all the recent logged events are show as “will happen” in 4 hours, that is, the logs are being saved either in a different time zone or they are not being interpreted correctly
05-15-2019 04:51 AM
Does anyone know if this issue has been resolved? Because on our servers, as well as on our client's servers, this is also an issue. Log time will show 2 hours earlier in our case
05-15-2019 05:37 AM
I thought this issue to be resolved and I dont personally see the symptoms you report here on my version of M.App Enterprise 2018 Update 3.
I suggest opening a support ticket for further investigation if issue perists for you.
05-21-2019 04:17 AM
Perhaps there is different MAE workflows showing different results here?
Upon further review I did find some discrepency, but I'm still unable to get the issue reported here either of 'in 4 hours' or 'in 3 hours' etc. I'm curious to know what is being done by user at time of these loggings of errors. Can you comment on that specifically or open a support ticket for further investigation please?
What I see on my side here just now is that I get an error and told to check log. When I do I see that it shows '11 minutes ago' when really it should say '1 minute ago'. So it seems like yes ok there is an issue here and I can log that. However, this is not the same as what others are reporting and agreeing with in this thread. For that I am unable to replicate. Perhaps it has something to do with region or time zone? I'm not certain at this point. But of course I will not argue against the numbers instead ask for more information so I can join the majority
So instead of logging a bug against what i see (we need to show a replicable case) and then hoping that any resolution to this will also resolve the issue from this discussion I would like to get more specific information to the case in an effor to replicate and report it. Preferably this can be done in support ticket, but if the information is provided in community and replicable I can be the one to open the ticket on user behalf.
Here is what I see now with Update 3 with just happened Studio error
Thank you for your patience with this matter and your assistance in providing additional information so that we can replicate this and move on to get it resolved permanently is greatly appreciated.
05-21-2019 04:25 AM
I'm also under the impression that it must have something to do with the timezone. Here in South Africa, we are GMT +2. As soon as an log is created, it will display as "2 hours ago". Again, this is in South Africa GMT +2.
Maybe the other users can also confirm this?
05-21-2019 02:37 PM - edited 05-21-2019 02:49 PM
This bug really puzzles me so I did some further testing... here are my findings:
I changed the mail configuration on the server to cause an error (i.e. server can´t send email to new user) got the same result log shows event “in 4 hours”. So it is not related to the workflow.
Went in and checked mapp_logmessage, noticed that the timestamp looks wrong, the stored time is 5 hours ahead of the local time, but the time zone is correct ( -5). Let’s say that I check the logs one second after the error was logged, if you make the math and subtract current time from the logged timestamp, the event will take place in 4 hours and 59 and 59 seconds minutes from now, that explains the “in 4 hours” message.
Next thing I tried was to change the server´s time zone and caused an error to see what happened. I expected it to show some difference, but nothing changed! Timestamp was still off by the same amount. Timezone still showed as -5 but that’s no surprise since postgres timezone is a session variable and does not change when the server´s timezone is modified
The conclusion that I get from this is that at some point MAE is sending the timestamp as UTC or wothout a timezone and it is not doing the conversion to the DB´s nor server´s timezone.
Here I´ll risk another assumption: If you try to reproduce the bug from a server located at Dublin, Edinburg, London and other UTC +00:00 location or if your database time zone is configured as UTC the bug can not be reproduced.
Workaround: As a way to 'fix' the bug you can change the postgres timezone by going into the “postgres.conf” file and changing the timezone variable to “UTC”. Then, after restarting postgres the logs will show fine in MAE.