Debian's cyrus 2.5.10 ctl_mboxlist segfault

Hello,

I'm using debian's cyrus 2.5.10 on a big installation for my students.  
On 26 Jully /run/cyrus, a tmpfs, became full
and cyrus start generate many such error :

2017-07-26T18:11:24.268202+02:00 cyrus-backend-etud-01  
cyrus/imap[22644]: autocreateinbox: User d21702882, INBOX was su
ccessfully created
2017-07-26T18:11:24.316643+02:00 cyrus-backend-etud-01  
cyrus/imap[22644]: autocreateinbox: User d21702882, subfolder Tr
ash creation succeeded.
2017-07-26T18:11:24.322689+02:00 cyrus-backend-etud-01  
cyrus/imap[22644]: autocreateinbox: User d21702882, subscription
    to Trash succeeded
2017-07-26T18:11:24.367002+02:00 cyrus-backend-etud-01  
cyrus/imap[22644]: autocreateinbox: User d21702882, subfolder Te
mplates creation succeeded.
2017-07-26T18:11:24.368849+02:00 cyrus-backend-etud-01  
cyrus/imap[22644]: autocreateinbox: User d21702882, subscription
    to Templates succeeded
2017-07-26T18:11:24.416869+02:00 cyrus-backend-etud-01  
cyrus/imap[22644]: autocreateinbox: User d21702882, subfolder Se
nt creation succeeded.
2017-07-26T18:11:24.418801+02:00 cyrus-backend-etud-01  
cyrus/imap[22644]: autocreateinbox: User d21702882, subscription
    to Sent succeeded
2017-07-26T18:11:24.420601+02:00 cyrus-backend-etud-01  
cyrus/imap[22644]: autocreateinbox: User d21702882, subfolder Dr
afts creation failed. System I/O error
2017-07-26T18:11:24.422464+02:00 cyrus-backend-etud-01  
cyrus/imap[22644]: autocreateinbox: User d21702882, subfolder sp
am creation failed. System I/O error
2017-07-26T18:11:24.422472+02:00 cyrus-backend-etud-01  
cyrus/imap[22644]: User d21702882, Inbox subfolders, created 3,
subscribed 3
2017-07-26T18:11:24.422487+02:00 cyrus-backend-etud-01  
cyrus/imap[22644]: created proc directory
2017-07-26T18:11:24.422520+02:00 cyrus-backend-etud-01  
cyrus/imap[22644]: IOERROR: creating /run/cyrus/proc/22644.new: No  
space left on device

At this point cyrus can't anymore start any new process.

The server has been reboot on July 30 but cyrus doesn't want to start  
anymore :

2017-07-30T11:30:19.550254+02:00 cyrus-backend-etud-01  
cyrus/ctl_cyrusdb[2062]: skiplist: clean shutdown file missing,  
updating recovery stamp
2017-07-30T11:30:19.550695+02:00 cyrus-backend-etud-01  
cyrus/ctl_cyrusdb[2062]: recovering cyrus databases
2017-07-30T11:30:56.139131+02:00 cyrus-backend-etud-01  
cyrus/ctl_cyrusdb[2062]: done recovering cyrus databases
2017-07-30T11:31:57.976574+02:00 cyrus-backend-etud-01  
cyrus/master[2045]: process type:START name:mupdatepush  
path:/usr/sbin/cyrus age:0.000s pid:2432 signaled to death by signal  
11 (Segmentation fault)
2017-07-30T11:31:57.976583+02:00 cyrus-backend-etud-01  
cyrus/master[2045]: can't run startup
2017-07-30T11:31:57.976586+02:00 cyrus-backend-etud-01  
cyrus/master[2045]: exiting

Back from hollidays I've tried serveral things :

     - /run/cyrus size has been double sized.
     - cyrus has been upgraded from 2.5.8 to 2.5.10
     - mailboxes.db has been converted to flat from twoskip with no  
success (still signal 11 on ctl_mboxlist -d and ctl_mboxlist -m).
     - reconstruct has been used on all mailboxes

flat mailboxes.db has 313716 lines

last lines of strace ctl_mboxlist -d :

stat("/var/lib/cyrus/mailboxes.db", {st_mode=S_IFREG|0600,  
st_size=37058416, ...}) = 0
fcntl(6, F_SETLKW, {type=F_UNLCK, whence=SEEK_SET, start=0, len=0}) = 0
fcntl(6, F_SETLKW, {type=F_RDLCK, whence=SEEK_SET, start=0, len=0}) = 0
fstat(6, {st_mode=S_IFREG|0600, st_size=37058416, ...}) = 0
stat("/var/lib/cyrus/mailboxes.db", {st_mode=S_IFREG|0600,  
st_size=37058416, ...}) = 0
fcntl(6, F_SETLKW, {type=F_UNLCK, whence=SEEK_SET, start=0, len=0}) = 0
--- SIGSEGV {si_signo=SIGSEGV, si_code=SEGV_MAPERR, si_addr=0x1c} ---
+++ killed by SIGSEGV +++

I don't know anymore what to do...

Sincerly,
         Jean Charles Delépine


----
Cyrus Home Page: http://www.cyrusimap.org/
List Archives/Info: http://lists.andrew.cmu.edu/pipermail/info-cyrus/
To Unsubscribe:
https://lists.andrew.cmu.edu/mailman/listinfo/info-cyrus