Re: [imapsync] Issue migrating large mailboxes from FirstClass to Dovecot

[Scott Barber <sabarber at me dot com>]

On Jul 12, 2010, at 5:12 PM, Phil Pearl (Lobbes) wrote:

> Hi Scott,
> 
>> OK, so the good news is that with the latest imapsync 1.321 and Mail::IMAPClient 3.25, the split/maxcommandlength stuff is working fine.
>> 
>> Now, I am at some point running into what seems like an issue in the reconnect logic.
> 
> What is the full command line you are using for imapsync?  Any chance
> you are specifying a timeout of 30?  I could speculate about a number
> of things but I'd like to get the full command line that you're using
> to get a little more big picture.  I would expect a timeout of 600s
> if you had the default settings.

That is correct, I am specifying a 30 second timeout against imapsync.  We saw similar results with the 600s timeout, figured no sense in waiting around for that long.

[... snip ... ]

>> Read:   * OK FirstClass IMAP4rev1 server v10.010 at xxxx.xxxx.xxx ready
> 
> This is a banner from a new/fresh connection...
> 
>> Sending: 2004 AUTHENTICATE CRAM-MD5
>> Read:   + PDIwMTAwNzA5MjEzODE1LjIzNDE5MjA0OEBsa3NkLm9yZz4=
>> Sending: Y2t1aG5zIDZhMWZmYzQ2Y2EyZjkyNDZjNjQ3NjY4ODQxMTAwZjc4
>> 

>> And now the reconnect fails -- I'm not sure why the authenticate fails, but perhaps we're not sending the userid again?
> 
> Based on the base64 encoded data here it appears we are sending a
> userid, are you seeing something different?

Whoops!  OK, missed that.  Thanks for pointing that out.

[ ... snip ... ]

>> I'm going to try this with 2.2.9 now that the reconnect logic is presumably fixed to see if we have the same problems there, but this latest set of issues has me a bit confused.
> 
> Any better luck with that?

Awaiting test data on this from the site.

Best,
Scott