Error interpreting the mail server response in Messenger Express
Background:
Migrated from Sparc running iMS 6.0 to Linux running iMS 6.3 (from CommSuite 5)
Problem:
Some users, some times experience problems using Messenger Express. They receive the following error in their web browsers: Error interpreting the mail server response.
Attempted solutions:
1) 120230-20 -- no change in behavior
2) reconstruct -r -- didn't fix the issue
3) Support case with Sun -- tech seems to be guessing as a solution and not trying to troubleshoot the issue (growing frustration -- those I am here)
Data gathered:
I turned on the telemetry logging for this user and in the end of the imap session, I see:
>0.002>)
F OK Completed
<30.747<C UNSELECT
>0.003>C OK Completed
NETWORK ERROR: Connection reset by peer
LOGOUT <username> 2007/7/17 8:46:07 - 2007/7/17 8:48:15
This seems to come and go (working sometimes and not working others) for a few (5 or so) users, though, I haven't found any correlation between the users or how they differ from anyone else. Checking email via POP or IMAP works fine (just not via Messenger express).
What else can I look at?
# 1
Hi,
I had a quick look at the code and the error you describe only seems to only occur once; when the webserver daemon (mshttpd) is trying to parse the response from the imap daemon (imapd) and there is a parsing error.
There should also have been an associated error in the http logs along the lines of "Failed to parse IMAP response at offset <something>". Does this ring any bells?
Regards,
Shane.
# 2
Thanks Shane.
the "Failed" error does seem to be happening in some cases, but not all from what I can tell. Here are a case in which the messages do appear:
General: Error reading from IMAP server: Timeout
General: imap::get_messagedata error at offset 2245. <and then header information>
General: imap::get_untagged error at offset 2245. <and then header information>
General: Failed to parse IMAP response at offset 2245: <and then header information>
Thoughts?
# 3
Hi,
Well it is entirely possible that you are hitting a known parsing error (bug #6569282 - mshttpd: Failed to parse IMAP response) which is scheduled to be fixed in the next messaging patch (120230-22) due at the end of this month or you are hitting a new parsing error - unfortunately without actually seeing the header data its hard to tell one way or another.
This bug is triggered by emails that have a specific header structure.
If you have a Sun support case open already they should be escalating to confirm whether you are hitting this particular bug or a new variant of this bug.
Regards,
Shane.
# 4
Shane...thank you so much. <refrains from bashing Sun Support at this point>
We have a valid support contract and the tech has never indicated that he has done any research into this issue -- he only keeps throwing out ideas based off of nothing. I asked him what he knows about the mentioned bug. awaiting response.
<pushing my luck> Is there anything I can provide you which would help diagnose this? Based off of the logs (snips i provided u before), it does indeed seem to be related to the email headers...specifcially headers SpamAssassin added.
Thanks again -- you're doing a great job.
Benji
# 5
Hi,
> Shane...thank you so much. <refrains from bashing Sun
> Support at this point>
I should point out that myself and Jay are also Sun Support (we answer forum questions in our spare time).
> We have a valid support contract and the tech has
> never indicated that he has done any research into
> this issue -- he only keeps throwing out ideas based
> off of nothing. I asked him what he knows about the
> mentioned bug. awaiting response.
I couldn't comment on this without looking at the case notes.
> <pushing my luck> Is there anything I can provide you
> which would help diagnose this? Based off of the
> logs (snips i provided u before), it does indeed
> seem to be related to the email
> headers...specifcially headers SpamAssassin added.
What's the Sun support case #? Have you provided the full logs (rather then just snips) to the support engineer?
> Thanks again -- you're doing a great job.
All part of the service.
Regards,
Shane.
# 6
Shane -- thanks again for answering. As a customer, I appreciate you being on this forum and Sun letting you be here. I think you have helped me several times with issues.
The case number is 65569739.
It is a little ironic that I found a mailing list post from you from several years ago -- apparently before you worked for Sun. You were asking about a Reply-To issue (which I also ran into).
Benji
# 7
Hi Benji,
The patch I mentioned earlier (120230-22) is now available.
I had a look at the case you mentioned (and added some comments to the Sun owner) but unfortunately I didn't see any notes in the case regarding additional debug information.
At this juncture your best bet is to get a copy and test/install -22 as this has a fix as I mentioned earlier for a bug matching the information provided thus far exactly and as an added bonus has few other useful bug-fixes (e.g. an annoying MTA queue expiration bug that has been impacting a few other customers).
Regards,
Shane.
# 8
Thanks for the update. it doesn't appear that 120230-22 is available to the public yet. Searching for 120230 provides me with 120230-20. When I search for 120230-22 specifically, it says it is restricted content. I will continue to watch for it and then apply it the first Monday morning (our maintenance window) after I download it.
Benji
