[Dovecot] Error from mobile client (Palm Pre)
Where I used to be able to access the server without incident, I can no longer receive mail on the mobile device. I'm seeing this in the server log:
Nov 27 23:45:05 bubba dovecot: imap-login: Login: user=<dmiller@amfes.com>, method=PLAIN, rip=173.117.196.198, lip=192.168.0.72, mpid=17422, TLS Nov 27 23:45:07 bubba dovecot: imap(dmiller@amfes.com): Error: FETCH [1.2] for mailbox INBOX UID 10652 got too little data: 0 vs 23553 Nov 27 23:45:07 bubba dovecot: imap(dmiller@amfes.com): Error: Corrupted index cache file /var/mail/amfes.com/dmiller/mdbox/mailboxes/INBOX/dbox-Mails/dovecot.index.cache: Broken MIME parts for mail UID 10652 Nov 27 23:45:07 bubba dovecot: imap(dmiller@amfes.com): Disconnected: FETCH failed bytes=88/4202
I HAVE made some tweaks to config - primarily adding plugins I never used before. These now include zlib and imap_zlib (I've tried removing imap_zlib - no change).
Daniel
On 11/27/2010 11:48 PM, Daniel L. Miller wrote:
Where I used to be able to access the server without incident, I can no longer receive mail on the mobile device. I'm seeing this in the server log:
Follow-up - only seeing this on one account - other accounts seem fine.
I'm deleting the config on the mobile and re-creating to see if it
resolves...I wonder if it's because an old message was saved on the
phone and the client is looking for it on the server - though why that
would cause such an error...
-- Daniel
On 11/28/2010 12:32 AM, Daniel L. Miller wrote:
On 11/27/2010 11:48 PM, Daniel L. Miller wrote:
Where I used to be able to access the server without incident, I can no longer receive mail on the mobile device. I'm seeing this in the server log:
Follow-up - only seeing this on one account - other accounts seem fine. I'm deleting the config on the mobile and re-creating to see if it resolves...I wonder if it's because an old message was saved on the phone and the client is looking for it on the server - though why that would cause such an error...
Ok - false alarm - it was indeed related to the client. Deleting & re-creating the account on the phone solved it. And re-enabling zlib & imap_zlib makes no apparent negative effect on the client - so I'll forget about it.
I do still want to know what caused the error...
Daniel
On 11/28/2010 1:01 AM, Daniel L. Miller wrote:
On 11/28/2010 12:32 AM, Daniel L. Miller wrote:
On 11/27/2010 11:48 PM, Daniel L. Miller wrote:
Where I used to be able to access the server without incident, I can no longer receive mail on the mobile device. I'm seeing this in the server log:
Follow-up - only seeing this on one account - other accounts seem fine. I'm deleting the config on the mobile and re-creating to see if it resolves...I wonder if it's because an old message was saved on the phone and the client is looking for it on the server - though why that would cause such an error...
Ok - false alarm - it was indeed related to the client. Deleting & re-creating the account on the phone solved it. And re-enabling zlib & imap_zlib makes no apparent negative effect on the client - so I'll forget about it.
I do still want to know what caused the error... And I spoke too soon again - I'm seeing this on other accounts using Thunderbird - so not client specific at the moment.
Daniel
On Sat, 2010-11-27 at 23:48 -0800, Daniel L. Miller wrote:
Nov 27 23:45:07 bubba dovecot: imap(dmiller@amfes.com): Error: FETCH [1.2] for mailbox INBOX UID 10652 got too little data: 0 vs 23553
You were using SIS, right? This looks kind of like it couldn't read the attachment. Can you fetch the message again later? For example:
doveadm fetch text mailbox INBOX uid 10652
On 11/29/2010 1:21 PM, Timo Sirainen wrote:
On Sat, 2010-11-27 at 23:48 -0800, Daniel L. Miller wrote:
Nov 27 23:45:07 bubba dovecot: imap(dmiller@amfes.com): Error: FETCH [1.2] for mailbox INBOX UID 10652 got too little data: 0 vs 23553 You were using SIS, right? This looks kind of like it couldn't read the attachment. Can you fetch the message again later? For example:
doveadm fetch text mailbox INBOX uid 10652
Yes, after digging into the man pages (nice job, by the way) I found that command. Using that, I was able to find that message, and delete it - which made the problem go away. For a moment - now I see it again on another account. I have not yet deleted THAT problem message. What can I do to further help diagnose this?
Also - can you do anything with doveadm to make it responsive to "Control-C"? Just so when a complete idiot, who looks a great deal like my reflection, types something like, "doveadm fetch -A text *" and hits a 20k+ mailbox, with attachments... I had to go in on another tty session to "kill -s 9 <doveadm-pid>"
Daniel
On 11/29/2010 6:36 PM, Daniel L. Miller wrote:
On 11/29/2010 1:21 PM, Timo Sirainen wrote:
On Sat, 2010-11-27 at 23:48 -0800, Daniel L. Miller wrote:
Yes, after digging into the man pages (nice job, by the way) I found that command. Using that, I was able to find that message, and delete it - which made the problem go away. For a moment - now I see it again on another account. I have not yet deleted THAT problem message. As usual - spoke too soon again. User deleted the message. I'm looking to see if any other accounts currently exhibit this fault.
-- Daniel
participants (2)
-
Daniel L. Miller
-
Timo Sirainen