#188 LoopDetection PHP FATAL

[cc-srforce]

Hey, we have this error that happens every time someone logs in via z-push on their phone:

30/01/2024 12:00:56 [1573651] [ERROR] [test1@ca-coeurdeflandre.fr] Please check your logs for this PID and errors like PHP-Fatals or Apache segmentation faults and report your results to the Z-Push dev team.
30/01/2024 12:00:59 [1556898] [ERROR] [test1@ca-coeurdeflandre.fr] LoopDetection->ProcessLoopDetectionPreviousConnectionFailed(): Command 'Sync' at 30/01/2024 12:00:56 with pid '1556898' terminated unexpectedly or is still running.
30/01/2024 12:00:59 [1556898] [ERROR] [test1@ca-coeurdeflandre.fr] Please check your logs for this PID and errors like PHP-Fatals or Apache segmentation faults and report your results to the Z-Push dev team.
30/01/2024 12:00:59 [1556898] [ERROR] [test1@ca-coeurdeflandre.fr] LoopDetection->ProcessLoopDetectionPreviousConnectionFailed(): Command 'Sync' at 30/01/2024 12:00:56 with pid '1556898' terminated unexpectedly or is still running.
30/01/2024 12:00:59 [1556898] [ERROR] [test1@ca-coeurdeflandre.fr] Please check your logs for this PID and errors like PHP-Fatals or Apache segmentation faults and report your results to the Z-Push dev team.
30/01/2024 12:02:09 [1556898] [ERROR] [test3@ca-coeurdeflandre.fr] LoopDetection->ProcessLoopDetectionPreviousConnectionFailed(): Command 'Sync' at 30/01/2024 12:00:31 with pid '1555411' terminated unexpectedly or is still running.
30/01/2024 12:02:09 [1556898] [ERROR] [test3@ca-coeurdeflandre.fr] Please check your logs for this PID and errors like PHP-Fatals or Apache segmentation faults and report your results to the Z-Push dev team.
30/01/2024 12:02:16 [1557752] [ERROR] [test3@ca-coeurdeflandre.fr] LoopDetection->ProcessLoopDetectionPreviousConnectionFailed(): Command 'Sync' at 30/01/2024 12:02:08 with pid '1556898' terminated unexpectedly or is still running.
30/01/2024 12:02:16 [1557752] [ERROR] [test3@ca-coeurdeflandre.fr] Please check your logs for this PID and errors like PHP-Fatals or Apache segmentation faults and report your results to the Z-Push dev team.
30/01/2024 12:02:16 [1557752] [ERROR] [test3@ca-coeurdeflandre.fr] LoopDetection->ProcessLoopDetectionPreviousConnectionFailed(): Command 'Sync' at 30/01/2024 12:02:09 with pid '1556898' terminated unexpectedly or is still running.
30/01/2024 12:02:16 [1557752] [ERROR] [test3@ca-coeurdeflandre.fr] Please check your logs for this PID and errors like PHP-Fatals or Apache segmentation faults and report your results to the Z-Push dev team.

I am currently using Backend 72 with the latest version of z-push

Thanks for your help

[cc-srforce]

After some research this error only appears with an Android phone by choosing the Exchange Active Sync authentication method present in Android with the "E-mail" application

The error does not appear with the same phone but when choosing the Exchange authentication method from Gmail

I also have no problem on an iPhone

And find this error in zpush error.log :

30/01/2024 13:33:48 [1591121] [WARN] [test1@ca-coeurdeflandre.fr] /usr/local/lib/z-push/backend/zimbra/zimbra.php:12762 date_default_timezone_set(): Timezone ID 'GMT+01.00/+02.00' is invalid (8)

(i've already installed the timezone.db extension from php) i'm using php-fpm7.4

[LiverpoolFCfan]

I believe the timezone in question is coming from an appointment. Was it originated from a system outside of zimbra?

if this is a timezone that appears on multiple appointments you can add it to the file zimbraNonPhpTimezones.php with a mapping to a normal php timezone.

If you are in France it would probably mean adding

$tzLookupList['GMT+01.00/+02.00'] = 'Europe/Paris';

to the list in the file.

Note: You will still get the warning forr the invalid timezones but the appointment should sync to the device with a proper timezone

[LiverpoolFCfan]

You will get these errors when sync is starting because Android starts a new ping after each newly synced folder but the previous one is still running. At thee end of the 60 (or whatever you have set) second ping cycle the older ones will be timed out by z-push and it will log that error message. Once all items are initially synced then you should rarely see the errors.

[cc-srforce]

Thanks for the answer,

I actually added the $tzLookupList in my zimbraNonPhpTimezones.php so it's good for this issues

However for the second error, if I take an agent from my structure (who is not the only one with so many logs)

I take the zpush.log from 12:00 p.m. on 01/30/2024 there are a lot of error logs while there are very few files to synchronize, I am attaching the zmmailbox gaf of the user as well as zpush logs

Thank you for your responsiveness

[LiverpoolFCfan]

So, the timing of the entries is generally spaced out which suggests you are getting proper ping cycles. Have you made sure to allow sessions of up to 1hr on your web server/proxy. These pings are considerably longer than your typical web request.