Still an issue: stuck processes

[Sebastian Hagedorn]

--On 5. Juli 2005 10:48:21 Uhr -0400 Ken Murchison <ken at oceana.com> wrote:

> Is this only a problem with pop3d or with imapd as well?

Both. There is a difference, though. At least some of the imadp processes 
seem to be stuck in a different spot:

(gdb) where
#0  0x0043dc88 in ___newselect_nocancel () from /lib/tls/libc.so.6
#1  0x08097545 in idle_notify ()
#2  0x0808f22d in mboxlist_findsub_alt ()
#3  0x0804fab9 in shut_down ()
#4  0x0804f455 in ?? ()
#5  0x08e12c48 in ?? ()
#6  0x08e12818 in ?? ()
#7  0x00000000 in ?? ()
(gdb)

(gdb) where
#0  0x001dff0e in __read_nocancel () from /lib/tls/libc.so.6
#1  0x00a67427 in BIO_new_socket () from /lib/libcrypto.so.4
#2  0x00a653e2 in BIO_read () from /lib/libcrypto.so.4
#3  0x00438c30 in ssl3_alert_code () from /lib/libssl.so.4
#4  0x00438dcc in ssl3_alert_code () from /lib/libssl.so.4
#5  0x0043a0cf in ssl3_read_bytes () from /lib/libssl.so.4
#6  0x0043affc in ssl3_get_message () from /lib/libssl.so.4
#7  0x00430cab in ssl3_accept () from /lib/libssl.so.4
#8  0x00430944 in ssl3_accept () from /lib/libssl.so.4
#9  0x0043fcaa in SSL_accept () from /lib/libssl.so.4
#10 0x0043b80d in ssl23_get_client_hello () from /lib/libssl.so.4
#11 0x0043b712 in ssl23_accept () from /lib/libssl.so.4
#12 0x0043fcaa in SSL_accept () from /lib/libssl.so.4
#13 0x0806d253 in index_expungeuidlist ()
#14 0x0805b011 in idle_update ()
#15 0x0804f485 in ?? ()
#16 0x00000000 in ?? ()
(gdb)

Of course I haven't checked every single one ... all of them (both pop3d 
and imapd) seem to have in common that they were connected to dial-up 
sessions - I can tell from the lsof output I get. Here's just one example:

imapd   27090 cyrus    0u  IPv4  498680173                TCP 
cyrus.rrz.uni-koeln.de:imaps->p50852077.dip0.t-ipconnect.de:1089 
(ESTABLISHED)
imapd   27090 cyrus    1u  IPv4  498680173                TCP 
cyrus.rrz.uni-koeln.de:imaps->p50852077.dip0.t-ipconnect.de:1089 
(ESTABLISHED)
imapd   27090 cyrus    2u  IPv4  498680173                TCP 
cyrus.rrz.uni-koeln.de:imaps->p50852077.dip0.t-ipconnect.de:1089 
(ESTABLISHED)

I'm guessing that the connection is lost and that's when the processes get 
stuck. I seem to remember that that's the scenario that led to the 
deadlocks previously.

>  I can't
> reproduce your problem here.

Well, I wouldn't know how to reproduce this in the first place ... except 
for the dropping dial-up connections.

>  Is there some kind of proxy or webmail
> process which might be unfriendly?

We use IMP for webmail, but the connections definitely don't come from 
there. We don't use POP or IMAP proxies.

Thanks, Sebastian Hagedorn
--
Sebastian Hagedorn - RZKR-R1 (Gebäude 52), Zimmer 18
Zentrum für angewandte Informatik - Universitätsweiter Service RRZK
Universität zu Köln / Cologne University - Tel. +49-221-478-5587
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 186 bytes
Desc: not available
Url : https://lists.andrew.cmu.edu/mailman/private/info-cyrus/attachments/20050705/c5bdcd6e/attachment.bin