[SOLVED] 2 Week Old Sync Problem with Outlook

[matttux]

Hi,

I have a client using Outlook 2007 with ZCO 5.0.16. About 2 weeks ago it stoped syncing. They still recieve everything from the server but anything that they create in outlook will not sync onto the server.

I have just being notified about this today and am wondering what is the best way to resync everything so all the sent mail, calendar, contacts created on outlook over the last two weeks will be back on the server?

I have tried to repair the connector but it will still not sync. Is there a way to force a resync from outlook back up to the server?

Thanks for any help.

[fsiegel]

There is probably a bad change record in there causing the problem. One thing might be to do put ZCO logging on, try one sync, and then exit. Then send the log. Perhaps there is a bad message that is forcing the sync to crash. Perhaps if we can identify the offending message, that would help.

[matttux]

Thanks for the reply. I have done that and attached the logs. Would it be worth updating the sync token to try get it to resync or am i likely to just cause more problems?


zimbra_log.txt
zimbra_log2.txt

[fsiegel]

Actually, neither of these logs shows me anything. I don't see that any sync was tried. Did you hit Send/Receive? Or more likely, your files got truncated. They only span 4 seconds, and they are both only 32K. Very suspicious. Perhaps you need to post them another way.

Also, I don't need the Search Protocol Host file -- just the "zimbra_log2.txt". And, you should turn calendar tracking off -- that just tends to mess things up and kill performance. That's only used for very specialized debugging. Just enable Logging and Http Logging. That should be enough.

So please try again, and make sure you try to do a sync.

I don't think updating the sync token will help you if you've got a bad change record in there. Let's see what the problem is first.

Thanks

[matttux]

I'm sorry about the log file, i've uploaded it again. Many thanks for the help

zimbra-log.log

[BillDean]

We have tried using the connector on our campus but we are beginning to move away from it because we are getting quite a few that end up having this same problem. The only fix seems to be to uninstall the connector, delete the zimbra cache files (saved where .pst files are also saved), and then reinstall and setup the connector again.

Unfortunately, our campus has no way of doing this remotely and we have to touch every machine where the connector fails, so we are trying to get our users to mainly use the web interface or the Zimbra Desktop application.

If anyone else knows how to keep this sort of thing from happening, please let us know.

[fsiegel]

Well, from Mattux's log files, here are the two most telling lines:

Change record required file upload of 98072723 bytes

ERROR: Zimbra::Rpc::Connection::WriteToHttp - failed on WinHttpWriteData(...), Error Code: 12030

It looks like there is some huge 98M appointment with an attachment that is timing out the connection. Unfortunately Http logging doesn't look like it's on, so I can't tell what the appt is, but there is probably some huge appointment trying to sync from Outlook that is the issue. Perhaps you can get rid of this and see if it helps.

There are tons of change records here. Is it possible just to create a new profile and initial sync again? I know that your issue is not to lose changes you have made in Outlook. Perhaps you can export the relevant Outlook folders to a .pst so you won't lose them, and see what you can merge back in? You could even use the pst import wizard to get those to Zimbra.

But first it would be good to identify the large appt on the Outlook side that is causing the problems. Also what is your maxmtamessage size?

[matttux]

Thanks very much for that fsiegel.

I turned on http logging and was able to find the offending email. The actual attachment was 2.32MB so not sure why it was reporting 98MB.

Once i removed it everything synced perfectly. Thanks for all the help, saved me a lot of head scratching.

[fsiegel]

Glad it worked out. I'm also not sure why the 98M, but that's what it seemed to think the size of the upload stream was. You did have a ton of change records in there, maybe from all those failed tries? Let me know if you have other issues.