It’s unlikely that even in NATted installation your public hostname resolves to 127.0.0.1. It should always point to the eth0 IP because Zimbra binds certain port to that address, and if it’s localhost you will have certain services available only in loopback.
DoSFilter logs to mailbox.log, you should check there. But in case you got trapped here you would have noticed errors in z-push-error.log.
About the broken object I don’t think it’s really an issue. May happen, but if it’s logged it should have been taken care (or just ignored). If it happens often it could possibly slow down sync, because instead of sending a block of object it has to try and retry with smaller chunks, but I don’t think this could mean a delay of minutes or hours.
I’ve no more ideas, you try creative ideas to figure it out :)
thanks for the links. I found the first one when googling, but it covers nginx and not apache. I was unsure how to translate nginx options like “fastcgi_read_timeout 3660” to apache syntax (I decided to translate it to “Timeout 3660” in the vhost section).
What iPhone setting exactly are you referring to?
However, Z-Push only sends new or modified or deleted items to the phone. If they are there, Z-Push’s job is done. The mobile device then notifies (or not) the user if there’s a new item.
@BerndHaensler said in No Notifications in iOS 12:
@Manfred : Not for all calendar, only for Shared calendars.
Is it your private calendar you share with another nextcloud user? Or is it someone else’s calendar you configured on your phone? In that case how did you add it to your phone? What is the exact setup (also on nextcloud)?
status update: changed from Xperia Z5 Compact (Android Nougat) to Samsung Galaxy Xcover 4 (Android Oreo 8.1). Works like a charm. No problems so far whatsoever. Most email clients seem to use the same system services to do ActiveSync, so they all had the same glitch, with one exception: nine, as mentioned before. they may have their own implementation. the only thing that concerns me is, older z-push versions have been working on my old phone.
the caldav backend is a community contribution and I’m not very familiar with it. I also don’t speak Korean, so I’ve no idea what ‘^@쨀\횛체쨩m짯 ^@\횚^@횋횥횂’ means or if it even makes sense. I have also no idea what the mobile packs into the timezone description which is in the incoming timezone tag.
I guess though that WBXML timezone and packed are different because one contains “Korea Standard Time” (’^@쨀\횛체쨩m짯 ^@\횚^@횋횥횂’) and another “(GMT+09:00) Seoul”. And in PHP it’s “Asia/Seoul”.
The timezone handling is tricky because PHP timezones doesn’t match Windows timezones, so you have to map them which can lead to inaccuracies.
Could you explain your use case more detailed? Does it involve only Outlook syncing to some groupware or do you also want to sync mobile devices?
@liverpoolfcfan thank you very much for the detailed explaination.
This server is the busiest we have, there are ~90 users and 130 AS sessions (not all active). Most of them are with normal handeld devices, but there is a percentage of Outlook clients, which are the most problematic ones.
Anyway the server load is very low, always lower than 0.3 and rarely going higher. Using ZWC is fast, so I don’t understand why it’s sometimes slow on AS replies. The infrastructure is not really up to date: Zimbra 8.6, PHP 5.5, ZP 2.3.9, backend 67.
I’m going to migrate state files to MySQL to see if it helps in improving performance.
@cjm51213 said in Overriding URLs:
… [#unknown] Zimbra->GetUserInfo(): Overriding Logon URL [https://mail.tclc.org/service/soap/] with ZIMBRA soapURL [https://mail.tclc.org:8443/service/soap/]
… [#unknown] Zimbra->GetUserInfo(): Overriding ZIMBRA_URL [https://mail.tclc.org] from config.php with ZIMBRA publicURL [https://mail.tclc.org:8443]
You should check your zimbra configuration and proxy configuration.
z-push will pick up whatever is set in your zimbra configuration for your domain under The General Information tab - Public service host name/protocol/port
Note: z-push should be pointed at the nginx proxy - not directly at a mailbox server - as connecting through the proxy is required in order to establish and maintain sessions.
Random musing: If the client machine is Windows 7 w/ Outlook 2016, it will attempt to use TLSv1.0 for AutoDiscover. This will be a problem if the webserver has disabled TLSv1.0 (for compatibility with PCI/DSS, for example). The rest of the system will connect and work fine. This is NOT a problem with Windows 10. The solution for Windows 7 is to make sure you have the TLSv1.1 and TLSv1.2 libraries installed, and make the registry patches to ensure that the stronger TLS libraries are used.
@markvnl thanks for your report. I tried adding that line and it seems to do something. Sadly after I enter authentication information I just get a blank page.
But this could be related to some other problem, at least I’m not asked again for u/p, which probably means they’ve been received.
what you have to do is to implement IBackend and ISearchProvider interfaces like the Kopano backend does which doesn’t extend the DiffBackend. Your exporter class (or your backend if you want to keep it in one place) has to implement IExportChanges interface (changes in your system to mobile). Your importer class (or your backend if you want to keep it in one place) has to implement IImportChanges interface (changes on the mobile to your system).
BackendCombined and BackendKopano might give you some ideas for the exporter/importer usage and how to move away from the DiffBackend.