El 15/02/2019 a las 8:47, Hajo Locke via dovecot escribió:
Hello,
Am 11.02.2019 um 09:12 schrieb Hajo Locke via dovecot:
Hello,
Am 08.02.2019 um 09:25 schrieb Hajo Locke via dovecot:
Hello List,
i have a problem with index-files which is separated in 2
subproblems. May be these problems are connected.
Currently we use Ubuntu 18.04 LTS which is bundled with dovecot
2.2.33.2
These servers are fresh installed machines and users are added to
the system after, there was no upgrade.
Sometimes it happens, that dovecot stops showing new mail. There is
no error in log, dovecot just seems to do his normal operations but
is not delivering new emails from /var/mail/myuser
emails are delivered again if i delete index-files from .imap
folder. So index is recreated and everything works again. I cant
reproduce this problem, but every 2 days i have one user who is
reporting this problem.
Iam surprised that no one has an opinion in this case. Nobody else
noticed this problem?
Yesterday i had a customer who was affected twice. First time we saw
an errormessage, which leads to an index problem:
Feb 13 15:46:58 hostname dovecot[30097]: imap(username): Error:
Corrupted record in index cache file
/home/popuser/username/mail/.imap/INBOX/dovecot.index.cache: UID
26876: Broken physical size in mailbox INBOX: read(/var/mail/username)
failed: Cached message size smaller than expected (2118 < 8088,
box=INBOX, UID=26876)
Feb 13 15:46:58 hostname dovecot[30097]: imap(username): Error:
read(/var/mail/username) failed: Cached message size smaller than
expected (2118 < 8088, box=INBOX, UID=26876) (FETCH BODY[] for mailbox
INBOX UID 26876)
At this point no new incoming mail is delivered by imapd. We deleted
index files, dovecot reacreated them and all was working again.
a day later same user was affected again. But like all other cases
before there is no error or unusual logline. i turned mail_debug on,
but no further hint.
There are definitely new Mails in /var/mail/username but dovecot still
shows old state from cache. I deleted index-files again and
imap-client showed a lot of new unread mails.
In log i can see that people are active in imap-clients and managing
mails. If a new mail is coming in this period it seems to harm
index-files some times. We never noticed same behaviour with long used
dovecot 2.2.22
It seems that 2.2.33.2 which is bundled with Ubuntu18.04 LTS has a
special problem.
Please help. What is your advice?
Second problem is similiar. I Upgraded a server from Ubuntu 16.04 to
Ubuntu 18.04. This includes an upgrade from dovecot 2.2.22 to
2.2.33.2. After that i installed an self packaged dovecot 2.2.36.1
Now here is same problem, dovecot is not showing new mails, but the
difference is we have a Raw backtrace in Log just as expected:
i think this crash is a special problem of dovecot 2.2.36.1.
if i downgrade to 2.2.33.2 all is working well, upgrading again to
2.2.36.1 leeds to same crash.
Feb 8 08:45:37 hostname dovecot[14882]: imap(myuser): Error: Raw
backtrace: /usr/lib/dovecot/libdovecot.so.0(+0xa1ee2)
[0x7f78b3b2cee2] -> /usr/lib/dovecot/libdovecot.so.0(+0xa1fda)
[0x7f78b3b2cfda] -> /usr/lib/dovecot/libdovecot.so.0(i_fatal+0)
[0x7f78b3abb8a8] ->
/usr/lib/dovecot/libdovecot-storage.so.0(index_mail_parse_header+0x78a)
[0x7f78b3e7940a] -> /usr/lib/dovecot/libdovecot.so.0(+0x83592)
[0x7f78b3b0e592] ->
/usr/lib/dovecot/libdovecot.so.0(i_stream_read+0x74)
[0x7f78b3b38814] ->
/usr/lib/dovecot/libdovecot.so.0(i_stream_read_data+0x3d)
[0x7f78b3b392fd] ->
/usr/lib/dovecot/libdovecot-storage.so.0(index_mail_get_header_stream+0x4a)
[0x7f78b3e7a2ea] ->
/usr/lib/dovecot/libdovecot-storage.so.0(mail_get_header_stream+0x4a)
[0x7f78b3dfd8da] ->
/usr/lib/dovecot/libdovecot-storage.so.0(imap_msgpart_open+0x24a)
[0x7f78b3ebb57a] -> dovecot/imap [myuser ip.ip.ip.ip
FETCH](+0x2081c) [0x5581ac76d81c] -> dovecot/imap [m038422
ip.ip.ip.ip FETCH](+0x1ea54) [0x5581ac76ba54] -> dovecot/imap
[myuser ip.ip.ip.ip FETCH](imap_fetch_more+0x39) [0x5581ac76cd69] ->
dovecot/imap [myuser ip.ip.ip.ip FETCH](cmd_fetch+0x31b)
[0x5581ac75e1bb] -> dovecot/imap [myuser ip.ip.ip.ip
FETCH](command_exec+0x5c) [0x5581ac76a09c] -> dovecot/imap [myuser
ip.ip.ip.ip FETCH](+0x1b612) [0x5581ac768612] -> dovecot/imap
[myuser ip.ip.ip.ip FETCH](+0x1b6ac) [0x5581ac7686ac] ->
dovecot/imap [myuser ip.ip.ip.ip FETCH](client_handle_input+0x18d)
[0x5581ac768a6d] -> dovecot/imap [myuser ip.ip.ip.ip
FETCH](client_input+0xac) [0x5581ac768fbc] ->
/usr/lib/dovecot/libdovecot.so.0(io_loop_call_io+0x52)
[0x7f78b3b43482] ->
/usr/lib/dovecot/libdovecot.so.0(io_loop_handler_run_internal+0x12e)
[0x7f78b3b44b9e] ->
/usr/lib/dovecot/libdovecot.so.0(io_loop_handler_run+0x36)
[0x7f78b3b43516] ->
/usr/lib/dovecot/libdovecot.so.0(io_loop_run+0x38) [0x7f78b3b436c8]
-> /usr/lib/dovecot/libdovecot.so.0(master_service_run+0x13)
[0x7f78b3ac6aa3] -> dovecot/imap [myuser ip.ip.ip.ip
FETCH](main+0x329) [0x5581ac75b159] ->
/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xe7)
[0x7f78b36bbb97] -> dovecot/imap [myuser ip.ip.ip.ip
FETCH](_start+0x2a) [0x5581ac75b2fa]
Feb 8 08:45:37 hostname dovecot[14882]: imap(myuser): Fatal:
master: service(imap): child 14135 killed with signal 6 (core dumped)
Thanks,
Hajo
Thanks,
Hajo
Hi,
A few days ago I posted a similar issue. My environment has RHEL6 and 7
servers. When I last updated the servers RHEL6 servers mantained
2.2.10-1_14.el6.x86_64 version, while RHEL7 updated dovecot from
2.2.10-8.el7.x86_64 to 2.2.36-3.el7.x86_64. When the RHEL7 servers (used
for sympa) processed a message for a user, its indexes were corrupted,
and the user could't access his inbox through webmail, so I had to
delete dovecot.* files from the user mail path to get it working again.
This is the error shown at the dovecot log:
Feb 5 09:33:25 buzon1 dovecot: imap(user@domain): Panic: file
mail-index-sync-keywords.c: line 227 (keywords_update_records):
assertion failed: (data_offset >= sizeof(struct mail_index_record))
Feb 5 09:33:25 buzon1 dovecot: imap(user@domain): Error: Raw backtrace:
/usr/lib64/dovecot/libdovecot.so.0() [0x3713268b8a] ->
/usr/lib64/dovecot/libdovecot.so.0() [0x3713268bf6] ->
/usr/lib64/dovecot/libdovecot.so.0() [0x37132224aa] ->
/usr/lib64/dovecot/libdovecot-storage.so.0(mail_index_sync_keywords+0x7fd)
[0x3713aca43d] ->
/usr/lib64/dovecot/libdovecot-storage.so.0(mail_index_sync_record+0xec)
[0x3713acacac] ->
/usr/lib64/dovecot/libdovecot-storage.so.0(mail_index_sync_map+0x234)
[0x3713acbae4] ->
/usr/lib64/dovecot/libdovecot-storage.so.0(mail_index_map+0x83)
[0x3713abcce3] ->
/usr/lib64/dovecot/libdovecot-storage.so.0(mail_index_refresh+0xe)
[0x3713ab793e] ->
/usr/lib64/dovecot/libdovecot-storage.so.0(maildir_sync_header_refresh+0x10)
[0x3713a4e1e0] -> /usr/lib64/dovecot/libdovecot-storage.so.0()
[0x3713a4e330] -> /usr/lib64/dovecot/libdovecot-storage.so.0()
[0x3713a4f3d4] -> /usr/lib64/dovecot/libdovecot-storage.so.0()
[0x3713a4f7b3] ->
/usr/lib64/dovecot/libdovecot-storage.so.0(maildir_storage_sync_init+0xd9)
[0x3713a4fa59] ->
/usr/lib64/dovecot/libdovecot-storage.so.0(mailbox_sync_init+0x31)
[0x3713a7d731] ->
/usr/lib64/dovecot/libdovecot-storage.so.0(mailbox_sync+0x27)
[0x3713a7e7b7] ->
/usr/lib64/dovecot/libdovecot-storage.so.0(index_storage_get_status+0x62)
[0x3713aa8fc2] -> /usr/lib64/dovecot/lib10_quota_plugin.so(+0xc3ec)
[0x7fd272dfe3ec] ->
/usr/lib64/dovecot/libdovecot-storage.so.0(mailbox_get_status+0x5c)
[0x3713a7f4fc] -> dovecot/imap(imap_status_get+0x7a) [0x41dbea] ->
dovecot/imap(cmd_status+0x179) [0x413059] ->
dovecot/imap(command_exec+0x3d) [0x4170bd] -> dovecot/imap() [0x416180]
-> dovecot/imap() [0x41627a] -> dovecot/imap(client_handle_input+0x11d)
[0x4164ed] -> dovecot/imap(client_input+0x6f) [0x41685f] ->
/usr/lib64/dovecot/libdovecot.so.0(io_loop_call_io+0x36) [0x3713278a56]
-> /usr/lib64/dovecot/libdovecot.so.0(io_loop_handler_run+0xa7)
[0x3713279b27] -> /usr/lib64/dovecot/libdovecot.so.0(io_loop_run+0x38)
[0x37132789c8]
Feb 5 09:33:25 buzon1 dovecot: imap(user@domain): Fatal: master:
service(imap): child 2057 killed with signal 6 (core dumped)
My solution was to downgrade dovecot and dovecot-pigeonhole back to
2.2.10-8.el7.x86_64
Regards
--
*Gonzalo Palacios Goicolea*
U.T de Infraestructura de Equipos Centrales
/Tecnologías de la Información/
Universidad Autónoma de Madrid • Campus de Cantoblanco
<3D"http://www.uam.es/";>
Antes de imprimir este correo piense si es necesario.Cuidemos el
medioambiente.
