@manfred said in Syncronization repeatedly breaks on one device:

Hi Achim,

could you ask the user to remove the account from the device, then remove the device with z-push-admin and then add the account on the device again.

Manfred

Yes, I will do so tomorrow and gather fresh logs!

There is another observation I made with the current 2.3.9 build but I am not sure if it is related with the observations above:
Many of my recurring all day appointments got vanished on my mobile devices. Especially halfway through the year I am missing 80% of the appointments. The missing items are consistent over all clients. I tried a resync or reinstalled accounts on my Android an iPhone devices with no luck. Upgrading z-push to 2.4.0 beta from the pre-final repo fixed those issues.

This is then even weirder, because dpkg runs the same command in postinst. It’s exactly the same thing.

Were there messages that stuff was fixed or (obsolete) states deleted? This would be the only explanation, that there was so much stuff to fix so it took longer. The second time everything is OK and it’s fast as expected.
But even in this scenario, I can’t imagine what the script would be doing for 20 minutes…

Hi @odix67 ,

Z-Push has not been tested for quite a while with Zarafa, but should still continue to work. When also using the packages provided by the z-push repository then you actually do not need to specify the backend provider used. but to be on the safe side, you’d need to set it to backendkopano.

Meanwhile there have also been quite some performance improvements in Kopano itself, so an upgrade there would not hurt either. for upgrading you just need to be aware of https://kb.kopano.io/display/WIKI/Migrating+from+ZCP+7.1+or+earlier+to+Kopano+Core+8.5+or+later.

Hi Bob,

in ActiveSync there’s no such thing as default contact number. Z-Push is not aware if a phone number was set as default.

“Use as default” is relevant to the mobiles only (I’m also not sure if this is something android specific or you can set that on other type of mobiles like iphones or windows phone as well).

Manfred

Hi Manfred,

i already reinstalled the packages.

Regards
Richard

@businesstux No feasible solution was found.

That was the guide that I had been following. I’m not sure if my implementation was incorrect, but adding the line:

AliasMatch (?i)/Autodiscover/Autodiscover.xml "/usr/share/z-push/autodiscover/autodiscover.php"

to my Apache configuration did not work. Once I created the autodiscover.domain.com DNS entry, then it started working for mobile devices.

The Windows 10 Mail client still does not work, but I think this is caused by a bug as outlined here:
https://forum.kopano.io/topic/570/z-push-autodiscover-on-2-3-7-not-working-as-expected/8

EDIT: The Outlook 2016 client works with z-push autodiscover once I install the Kopano Outlook Extension. All is well!

Thank you!

Thank you - it’s not so easy :-(

I don’t know what to say… In the meantime i’ve reinstalled the Kopano Core, Kopano WebApp and Z-Push for Kopano. Now all users are working well. I think there was a problem at installation?

Before the reinstall i’ve made a snapshot of the virtual machine. But, what happened? After the rollback to the UCS with the old, not reinstalled Kopano, the same user, which had a problem with ActiveSync works now? Every mail, every contact, every calendar entry was synched.

I don’t believe in ghosts, until now. But there was only an update for the iOS Outlook app. I don’t have the exact update time, but the update was released 18 hours ago. I think the iphone have installed the update in the last 1-2 hours and MS fixed a bug in relation to ActiveSync. That’s my only explanation for the problem. That would explain the successful test with the testuser.

Thanks for the replys.

Robin

For all with such a problem:
i’ve got a script via subscription support to pruge the items outside of kopanos triggers and internal actions.

its a script called “purge-softdelete.py”, and deletes the items directly from db.
seems to work in test environment - i will provide info if this is also for the productive env.

one thing to mention: availible ram is getting lower and lower within the process and needs some time (about 1d to release the ressources or requires rebooting the server)

//EDIT: A bit late, but here info for productive Systems:
It works and solved the problem, but NEVER EVER start the purge with 30 days or something if you didnt purge for over half a year and the script want to delete over 1,8 mio elements ;)
start with 360 days and decrase 30 days with every purge.
Ignore the MAPI_E_DISK_ERROR’s, the script is searching for them and will try to redo the action afterwards. if you reached your desired days run it 3 till 4 times to eleminate all items, even with MAPI_E_DISK_ERROR…
And to mention, dont run the initial purge if users are active, the performance of db will decrase and the users will blame you ;(

coffee_is_life

Also, thanks to you @manfred, Thanks for explanation. (It’s the only (i think) section in config.php file that isn’t documented at all in comments.)

@fbartels
Thanks a lot! finaly got it working!! spent hours of searching and it was this simple … :D

Interesting. That seems to have fixed it, thanks for the pointer. Strange that it only broke for us with the new version of Z-Push.

Hi Ralf,

thanks for the update. Good to hear that it works now.

Manfred

It also might depend on the number of syncs running in parallel. When switching to kopano, we always had several syncs running and a 10 GB mailbox took a night for the initial sync. Now I had to resync a 14 GB mailbox in normal operation and this took about 2 hours.

I use “Task & Notes for MS Exchange” UniqTec Inc.

for Android.

It’s actually great!

It costs a tiny amount of money, but I can use it on how many Android devices I want.

And it works perfectly well with Kopano z-push.