Thanks, Mark. The copy of the imapd being used was directly downloaded
from UW and thus should be unmodified.

I compared two cases -- one is using the "spam" button and the other
is the client's normal forward function, and found the following
difference in the termination message(var/log/syslog):

Case1: Using "Spam" button -- connection drop error message returned

Dec 26 10:10:14 poisson imapd[28495]: [ID 323218 mail.debug] tid= 1:
unlocking sessionLock

In contrast, in the case 2 that is using forward function ( the
operation successed without connection drop error message),  a lot
more mail info has been logged:

Dec 26 10:13:20 testmachine imapd[28515]: [ID 323218 mail.debug] tid=
1: unlocking s
essionLock
Dec 26 10:13:20 testmachine imapd[28515]: [ID 453631 mail.debug] tid=
1: Adding conn
ection (serverAddr=192.168.1.50)
Dec 26 10:13:20 testmachine imapd[28515]: [ID 816976 mail.debug] tid=
1: Connection
added [1]
Dec 26 10:13:20 testmachine imapd[28515]: [ID 467101 mail.debug] tid=
1: connectionI
D=1025
Dec 26 10:13:20 testmachine imapd[28515]: [ID 805042 mail.debug] tid=
1: shared=1
Dec 26 10:13:20 testmachine imapd[28515]: [ID 982078 mail.debug] tid=
1: usedBit=0
Dec 26 10:13:20 testmachine imapd[28515]: [ID 727660 mail.debug] tid=
1: threadID=1
Dec 26 10:13:20 testmachine imapd[28515]: [ID 577507 mail.debug] tid=
1: serverAddr=
192.168.1.50
Dec 26 10:13:20 testmachine imapd[28515]: [ID 939703 mail.debug] tid=
1: AuthType=1
Dec 26 10:13:20 testmachine imapd[28515]: [ID 142272 mail.debug] tid=
1: TlsType=0
Dec 26 10:13:20 testmachine imapd[28515]: [ID 537450 mail.debug] tid=
1: SaslMech=0
Dec 26 10:13:20 testmachine imapd[28515]: [ID 625532 mail.debug] tid=
1: SaslOpt=0
Dec 26 10:13:20 testmachine imapd[28515]: [ID 339871 mail.debug] tid=
1: hostCertPat
h=/var/ldap
Dec 26 10:13:20 testmachine imapd[28515]: [ID 639905 mail.debug] tid=
1: userID=uid=
tester,ou=People,dc=testdomain,dc=com
Dec 26 10:13:21 testmachine imapd[28515]: [ID 453631 mail.debug] tid=
1: Adding conn
ection (serverAddr=192.168.1.50)
Dec 26 10:13:21 testmachine imapd[28515]: [ID 816976 mail.debug] tid=
1: Connection
added [2]
Dec 26 10:13:21 testmachine imapd[28515]: [ID 467101 mail.debug] tid=
1: connectionI
D=1026
Dec 26 10:13:21 testmachine imapd[28515]: [ID 805042 mail.debug] tid=
1: shared=1
Dec 26 10:13:21 testmachine imapd[28515]: [ID 982078 mail.debug] tid=
1: usedBit=0
Dec 26 10:13:21 testmachine imapd[28515]: [ID 727660 mail.debug] tid=
1: threadID=1
Dec 26 10:13:21 testmachine imapd[28515]: [ID 577507 mail.debug] tid=
1: serverAddr=
192.168.1.50
Dec 26 10:13:21 testmachine imapd[28515]: [ID 939703 mail.debug] tid=
1: AuthType=1
Dec 26 10:13:21 testmachine imapd[28515]: [ID 142272 mail.debug] tid=
1: TlsType=0
Dec 26 10:13:21 testmachine imapd[28515]: [ID 537450 mail.debug] tid=
1: SaslMech=0
Dec 26 10:13:21 testmachine imapd[28515]: [ID 625532 mail.debug] tid=
1: SaslOpt=0
Dec 26 10:13:21 testmachine imapd[28515]: [ID 339871 mail.debug] tid=
1: hostCertPat
h=/var/ldap
Dec 26 10:13:21 testmachine imapd[28515]: [ID 639905 mail.debug] tid=
1: userID=uid=
tester,ou=People,dc=testdomain,dc=com
Dec 26 10:13:21 testmachine imapd[28515]: [ID 234311 mail.info] Login
user=tester hos
t=testmachine [192.168.1.23]
Dec 26 10:13:21 testmachine imapd[28515]: [ID 960700 mail.info] Logout
user=tester ho
st=testmachine [192.168.1.23]


I am not sure what we can conclude from the above differences.


Derek


On Dec 25, 2007 7:35 PM, Mark Crispin <[EMAIL PROTECTED]> wrote:
> Unless you have memory limits set ridiculously low (as in suitable for the
> early 1980s), it is quite unlikely that imapd would fail due to lack of
> memory.  If it did, there would be an entry in the mail syslog reading
> "IMAP toolkit crash: Out of memory".
>
> Nor is the IMAP command "FETCH 106 BODY[HEADER]" a particularly difficult,
> unexpected, or risky operation for an IMAP server to undertake.
>
> Rather than guess at the problem, and trying all sorts of things at random
> until you find the measure that fixes it, I suggest that you try a little
> bit more investigation first.
>
> As noted above, UW imapd leaves its log messages in the mail syslog.  [At
> least it does in unmodified UW imapd.  If you have a copy of UW imapd that
> was modified by a third party, the first thing to do is see if unmodified
> UW imapd directly from UW works any better.]  So, take a look in the mail
> syslog and see if there are any interesting log messages corresponding
> with the incident.
>
> UW imapd normally logs its startup, authentication, and termination.  The
> last is probably the most interesting.  Is there a termination message, or
> did that imapd vanish without a trace?  If there is a termination message,
> what did the message say?  That will point you towards the next step in
> your debugging.
>
>
> On Tue, 25 Dec 2007, Derek Xu wrote:
> > Hello --
> >
> > I am running UW IMAP 2007. The mail client is squirrel webmail 1.4.13
> > with spam_buttons plugin. When I click on the "spam" button, it always
> > generates the following error message:
> >
> > ERROR: Connection dropped by IMAP server.
> > Query: FETCH 106 BODY[HEADER]
> >
> > I queried the squirrel webmail mailing list, they suspected it is an
> > IMAP-related issue.  However, the mail log doesn't provide any useful
> > information. I just came across an article talking about bincimap and
> > squeirrelMail. There was a similar problem and the suggested solution
> > is use /usr/local/bin/softlimit to set memory resource,such as  stack
> > segment per process, physical pages per process and data segment per
> > process, for bincimap to  much higher values.
> >
> > http://osdir.com/ml/mail.imap.binc.general/2003-10/msg00089.html
> >
> > I wonder if there is any similar way to change the memory resource
> > for UW imapd. What normally cause imapd drop connections with clients?
> >
> > Thanks,
> >
> > Derek
> > _______________________________________________
> > Imap-uw mailing list
> > [email protected]
> > https://mailman1.u.washington.edu/mailman/listinfo/imap-uw
> >
>
> -- Mark --
>
> http://panda.com/mrc
> Democracy is two wolves and a sheep deciding what to eat for lunch.
> Liberty is a well-armed sheep contesting the vote.
>
_______________________________________________
Imap-uw mailing list
[email protected]
https://mailman1.u.washington.edu/mailman/listinfo/imap-uw

Reply via email to