I’m getting mad…
I’ve a new server using Zimbra and ZPush 2.4.4 with the latest available backend.
In Zimbra I’ve whitelisted the server public IP address, localhost addresses and the office IP.
Now a user changed his password in Zimbra and the mobile phones are not prompting for the new password and Z-Push keeps authenticating to Zimbra with wrong credentials. And despite of all the whitelist above (I’ve restarted all zimbra) I keep getting 503 Service unavailable from Zimbra!
Z-Push is configured directly in zimbra’s nginx.conf using this method.
The only strange thing I noticed in mailbox.log is that I get two IPs in the oip field. That is:
2018-09-12 14:22:29,398 INFO [qtp127618319-823:http://127.0.0.1/service/soap/] [name=username;oip=OFFICE_IP, 127.0.0.1;port=49930;ua=Android/8.0.0-EAS-2.0(...717940) devip=OFFICE_IP ZPZB/68;] SoapEngine - handler exception: authentication failed for [username], invalid password
I cannot understand why in oip I see the remote IP and the z-push server IP. Now Z-Push is configured with https://127.0.0.1 as Zimbra server (and without url override), but even if I use the public hostname I will see Zimbra’s public IP instead of 127.0.0.1.
It’s driving me mad, I cannot see why Zimbra keeps throttling Z-Push connections!