daemontools

Hi Roberto,

My system is Ubuntu 24.04.2 LTS with gcc version 13.3.0. and daemontools 0.81 has some issues.
The errors starts to appear in the logs after the upgrade to 0.81.

The times of the errors matches with these cron jobs:

0 0 * * * /usr/local/bin/svc -a /service/qmail-submission/log
0 0 * * * /usr/local/bin/svc -a /service/qmail-smtpd/log
0 0 * * * /usr/local/bin/svc -a /service/qmail-smtpsd/log
0 0 * * * /usr/local/bin/svc -a /service/qmail-send/log
0 0 * * * /usr/local/bin/svc -a /service/vpopmaild/log
0 0 * * * /usr/local/bin/svc -a /service/vusaged/log

Here is the errors in syslog and dmesg.

root@qmail:~# dmesg -T |grep multilog
[Sat Sep 13 00:00:02 2025] multilog[44646]: segfault at 58929c38753e ip 000058919c38086f sp 00007ffdb3455c20 error 4 in multilog[58919c37f000+5000] likely on CPU 0 (core 0, socket 0)
[Sat Sep 13 00:00:02 2025] multilog[44645]: segfault at 625a1e92853e ip 000062591e92186f sp 00007ffe308d52f0 error 4 in multilog[62591e920000+5000] likely on CPU 1 (core 0, socket 1)
[Sat Sep 13 00:00:02 2025] multilog[44651]: segfault at 565ea33a353e ip 0000565da339c86f sp 00007fff053139d0 error 4 in multilog[565da339b000+5000] likely on CPU 0 (core 0, socket 0)
[Sat Sep 13 00:00:02 2025] multilog[44631]: segfault at 5e3ea4d7353e ip 00005e3da4d6c86f sp 00007fff1ee57d90 error 4 in multilog[5e3da4d6b000+5000] likely on CPU 1 (core 0, socket 1)
[Sat Sep 13 00:00:02 2025] multilog[44652]: segfault at 64961fe5053e ip 000064951fe4986f sp 00007ffd41c9c8a0 error 4 in multilog[64951fe48000+5000] likely on CPU 1 (core 0, socket 1)
[Sat Sep 13 00:00:02 2025] multilog[44633]: segfault at 58ce60be153e ip 000058cd60bda86f sp 00007ffe903eee80 error 4 in multilog[58cd60bd9000+5000] likely on CPU 0 (core 0, socket 0)
[Sun Sep 14 00:00:04 2025] multilog[61005]: segfault at 619f6fd0353e ip 0000619e6fcfc86f sp 00007ffe9edfa800 error 4 in multilog[619e6fcfb000+5000] likely on CPU 1 (core 0, socket 1)
[Sun Sep 14 00:00:04 2025] multilog[60997]: segfault at 64556419153e ip 000064546418a86f sp 00007ffc1aeabbe0 error 4 in multilog[645464189000+5000] likely on CPU 0 (core 0, socket 0)
[Sun Sep 14 00:00:04 2025] multilog[61000]: segfault at 5c6e4497653e ip 00005c6d4496f86f sp 00007ffe28ce1710 error 4 in multilog[5c6d4496e000+5000] likely on CPU 1 (core 0, socket 1)
[Sun Sep 14 00:00:04 2025] multilog[61006]: segfault at 596228ab153e ip 0000596128aaa86f sp 00007ffe1211cab0 error 4 in multilog[596128aa9000+5000] likely on CPU 1 (core 0, socket 1)
[Sun Sep 14 00:00:04 2025] multilog[61003]: segfault at 60aa7e53b53e ip 000060a97e53486f sp 00007fff764fade0 error 4 in multilog[60a97e533000+5000] likely on CPU 1 (core 0, socket 1)
[Sun Sep 14 00:00:04 2025] multilog[61001]: segfault at 55b64fb5953e ip 000055b54fb5286f sp 00007ffe9a59d720 error 4 in multilog[55b54fb51000+5000] likely on CPU 1 (core 0, socket 1)
[Mon Sep 15 00:00:08 2025] multilog[392292]: segfault at 58e32d37053e ip 000058e22d36986f sp 00007ffedee869d0 error 4 in multilog[58e22d368000+5000] likely on CPU 0 (core 0, socket 0)
[Mon Sep 15 00:00:08 2025] multilog[392300]: segfault at 59d4111b553e ip 000059d3111ae86f sp 00007fffa74bb710 error 4 in multilog[59d3111ad000+5000] likely on CPU 1 (core 0, socket 1)
[Mon Sep 15 00:00:08 2025] multilog[392280]: segfault at 575a8510253e ip 00005759850fb86f sp 00007ffc528ca570 error 4 in multilog[5759850fa000+5000] likely on CPU 0 (core 0, socket 0)
[Mon Sep 15 00:00:08 2025] multilog[392284]: segfault at 6544da5d353e ip 00006543da5cc86f sp 00007ffdebbbeb00 error 4 in multilog[6543da5cb000+5000] likely on CPU 0 (core 0, socket 0)
[Mon Sep 15 00:00:08 2025] multilog[392298]: segfault at 64258f6a053e ip 000064248f69986f sp 00007ffeb1fb7ed0 error 4 in multilog[64248f698000+5000] likely on CPU 0 (core 0, socket 0)
[Mon Sep 15 00:00:08 2025] multilog[392296]: segfault at 60f0035ca53e ip 000060ef035c386f sp 00007ffc20777c60 error 4 in multilog[60ef035c2000+5000] likely on CPU 1 (core 0, socket 1)
[Tue Sep 16 00:00:11 2025] multilog[723068]: segfault at 57812c09053e ip 000057802c08986f sp 00007ffc0a4055e0 error 4 in multilog[57802c088000+5000] likely on CPU 1 (core 0, socket 1)
[Tue Sep 16 00:00:11 2025] multilog[723054]: segfault at 58b2a31d653e ip 000058b1a31cf86f sp 00007ffd1f270420 error 4 in multilog[58b1a31ce000+5000] likely on CPU 0 (core 0, socket 0)
[Tue Sep 16 00:00:11 2025] multilog[723061]: segfault at 60a4fa60753e ip 000060a3fa60086f sp 00007ffd1f1c19b0 error 4 in multilog[60a3fa5ff000+5000] likely on CPU 1 (core 0, socket 1)
[Tue Sep 16 00:00:11 2025] multilog[723067]: segfault at 5e19fcfe853e ip 00005e18fcfe186f sp 00007fff74b6a610 error 4 in multilog[5e18fcfe0000+5000] likely on CPU 0 (core 0, socket 0)
[Tue Sep 16 00:00:11 2025] multilog[723069]: segfault at 5bd8d114b53e ip 00005bd7d114486f sp 00007ffda756e5d0 error 4 in multilog[5bd7d1143000+5000] likely on CPU 1 (core 0, socket 1)
[Tue Sep 16 00:00:11 2025] multilog[723064]: segfault at 619622f9b53e ip 0000619522f9486f sp 00007fff3db93740 error 4 in multilog[619522f93000+5000] likely on CPU 1 (core 0, socket 1)
[Wed Sep 17 00:00:14 2025] multilog[1229085]: segfault at 594d71c5353e ip 0000594c71c4c86f sp 00007ffc963304d0 error 4 in multilog[594c71c4b000+5000] likely on CPU 1 (core 0, socket 1)
[Wed Sep 17 00:00:14 2025] multilog[1229073]: segfault at 644a3ca7e53e ip 000064493ca7786f sp 00007ffdd50f0100 error 4 in multilog[64493ca76000+5000] likely on CPU 0 (core 0, socket 0)
[Wed Sep 17 00:00:14 2025] multilog[1229076]: segfault at 611e231e053e ip 0000611d231d986f sp 00007fff4809c670 error 4 in multilog[611d231d8000+5000] likely on CPU 0 (core 0, socket 0)
[Wed Sep 17 00:00:14 2025] multilog[1229070]: segfault at 616531b5153e ip 0000616431b4a86f sp 00007ffd5c82ab10 error 4 in multilog[616431b49000+5000] likely on CPU 1 (core 0, socket 1)
[Wed Sep 17 00:00:14 2025] multilog[1229082]: segfault at 5df859c7a53e ip 00005df759c7386f sp 00007ffe07edad50 error 4 in multilog[5df759c72000+5000] likely on CPU 0 (core 0, socket 0)
[Wed Sep 17 00:00:14 2025] multilog[1229084]: segfault at 5fbf5d1db53e ip 00005fbe5d1d486f sp 00007fff599cf1d0 error 4 in multilog[5fbe5d1d3000+5000] likely on CPU 0 (core 0, socket 0)
[Thu Sep 18 00:00:17 2025] multilog[1724821]: segfault at 5d334bc4253e ip 00005d324bc3b86f sp 00007fff63bfb370 error 4 in multilog[5d324bc3a000+5000] likely on CPU 1 (core 0, socket 1)
[Thu Sep 18 00:00:17 2025] multilog[1724810]: segfault at 5f13f1c6553e ip 00005f12f1c5e86f sp 00007ffd649092f0 error 4 in multilog[5f12f1c5d000+5000] likely on CPU 1 (core 0, socket 1)
[Thu Sep 18 00:00:17 2025] multilog[1724805]: segfault at 5812ad49d53e ip 00005811ad49686f sp 00007ffcd364f840 error 4 in multilog[5811ad495000+5000] likely on CPU 1 (core 0, socket 1)
[Thu Sep 18 00:00:17 2025] multilog[1724806]: segfault at 5699de94c53e ip 00005698de94586f sp 00007fff91530530 error 4 in multilog[5698de944000+5000] likely on CPU 0 (core 0, socket 0)
[Thu Sep 18 00:00:17 2025] multilog[1724823]: segfault at 616a68d7253e ip 0000616968d6b86f sp 00007ffd003ac540 error 4 in multilog[616968d6a000+5000] likely on CPU 0 (core 0, socket 0)
[Thu Sep 18 00:00:17 2025] multilog[1724816]: segfault at 56fef349e53e ip 000056fdf349786f sp 00007fff931817e0 error 4 in multilog[56fdf3496000+5000] likely on CPU 0 (core 0, socket 0)
[Fri Sep 19 00:00:20 2025] multilog[2183285]: segfault at 6287f43b553e ip 00006286f43ae86f sp 00007fff7601c970 error 4 in multilog[6286f43ad000+5000] likely on CPU 1 (core 0, socket 1)
[Fri Sep 19 00:00:20 2025] multilog[2183273]: segfault at 6372d990d53e ip 00006371d990686f sp 00007ffed78a6c00 error 4 in multilog[6371d9905000+5000] likely on CPU 1 (core 0, socket 1)
[Fri Sep 19 00:00:20 2025] multilog[2183238]: segfault at 634c18fab53e ip 0000634b18fa486f sp 00007ffd0fa350f0 error 4 in multilog[634b18fa3000+5000] likely on CPU 0 (core 0, socket 0)
[Fri Sep 19 00:00:20 2025] multilog[2183254]: segfault at 580e6b29453e ip 0000580d6b28d86f sp 00007ffd3f7dcdb0 error 4 in multilog[580d6b28c000+5000] likely on CPU 0 (core 0, socket 0)
[Fri Sep 19 00:00:20 2025] multilog[2183278]: segfault at 60e21ef1c53e ip 000060e11ef1586f sp 00007ffd52b9bcc0 error 4 in multilog[60e11ef14000+5000] likely on CPU 1 (core 0, socket 1)
[Fri Sep 19 00:00:20 2025] multilog[2183212]: segfault at 653f7592853e ip 0000653e7592186f sp 00007ffef03145d0 error 4 in multilog[653e75920000+5000] likely on CPU 1 (core 0, socket 1)

Can you try this one please? https://github.com/sagredo-dev/daemontools/releases/tag/v0.82

Hi Roberto,

It seems the segfault issue in version 0.81 is fixed. I tested the new version and found no related multilog errors in syslog or dmesg.

Hi Bai,

I'm trying to stress the multilog rotation in order to debug a segfault but I still can get one..

daemontools - package/install - puts in rc.local : csh /command/svcscanboot & but the rc.local is driven by bash and csh is not installed by default on ubuntu - so it fails

thanks for the hint. Solved in v. 0.78.3

HI Roberto, 

I am facing issue while package/install of daemontools on RHEL6/CentOS6 and below is the error message 

./compile timestamp.c
./load multilog deepsleep.o timestamp.o match.o time.a unix.a byte.a
timestamp.o: In function `readable_datetime':
timestamp.c:(.text+0xc): undefined reference to `clock_gettime'
collect2: ld returned 1 exit status
make: *** [multilog] Error 1

Regards

Shailendra

Hi Shailendra,

that clock_gettime function is in the time.h file of the glibc library. Probably it is not linked to the timestamp.c program. Which glibc version do you have? (ldd --version)

Can you modify the src/conf-cc file like this

-gcc -O2
+gcc -lrt -O2

and run package/install again?

Anyway you can solve installing the original daemontools-0.76 patched with the errno.patch, but please let me know if the above test works for centos 6.10

Hi Roberto, 

Tried the hack that you suggested but still getting the same error

./compile timestamp.c
./load multilog deepsleep.o timestamp.o match.o time.a unix.a byte.a
timestamp.o: In function `readable_datetime':
timestamp.c:(.text+0xc): undefined reference to `clock_gettime'
collect2: ld returned 1 exit status
make: *** [multilog] Error 1

Kindly find the output of ldd --version

ldd (GNU libc) 2.12
Copyright (C) 2010 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
Written by Roland McGrath and Ulrich Drepper.

As suggested by you I tried with daemontools 0.76 and I get the following error

./load envdir unix.a byte.a
/usr/bin/ld: errno: TLS definition in /lib64/libc.so.6 section .tbss mismatches non-TLS reference in envdir.o
/lib64/libc.so.6: could not read symbols: Bad value
collect2: ld returned 1 exit status
make: *** [envdir] Error 1

Regards

Shailendra

Sorry for the wrong suggestion I gave before. Use conf-cc and conf-ld like these

conf-cc with no modifications, so

gcc -O2

add -lrt to conf-ld

gcc -s -lrt

I don't know what's going on for version 0.76

Hi Roberto, 

As per your suggesstion modifying the conf-ld with gcc -s -lrt did the trick and I was able to successfully install daemontools 0.78.1 on RHEL6/Centos6.

Thanks for your help

Regards

Shailedra

Hello Roberto

Is package/compile script missing from daemontools-0.78 source package?

BTW thanks for all the great work you and others are doing to keep qmail alive.

Unfortunately it is missing. I'm out of home now. I'll correct the problem tomorrow. 

In the meantime you can replace that script with one of a previous version

Many thanks Roberto and my heartfelt appreciation for the incredible work your are doing.

I've rebuilt the 0.78 package with the package/compile script. 

Thanks for the advice!

Hi Roberto,

package/compile script still missing in daemontools-0.78 source?

Tx Bai, I've created 0.78.1 with that file added

cd daemontools-​${DAEMONTOOLS_VER}

this line also has same 3 hidden chars 9see in hex)

Thank you. I retyped those string

cp -rp ${SRC}/daemontools-${DAEMONTOOLS_VER}/admin/daemontools-​${DAEMONTOOLS_VER} .

this line somehow has 3 hidden bytes (copy it and look in hex) which makes it wrong

is it an idea to put the following already here instead of on page installing and configuring vpopmail?:

apt install build-essential autoconf automake libmariadb-dev libmariadb-dev-compat

I created a page with the preinstallation tasks, where this and other preliminary tasks are covered. Feel free to post other hints 

ok, inserted

Hi everybody .

i had a problem need your help :

- when i run a command line :

# package/install 

i see : 

Linking ./src/* into ./compile...
Compiling everything in ./compile...
package/compile: 20: exec: make: not found . 

help me , pls .

thanks.

I think that you should install autotools. Please refer to your distribution's documentation for the purpose

Hello,

I´m doing a procedure to install qmail to CentOS 8 (systemd)

run file gets 644 attribute, should we change to 744 ?

Thanks for your work !

Marco Varanda

yes

If you receive error: Error supervise not running
It's because daemontools/svscanboot is not starting from sysvinit inittab.

To start from systemd:

cat > /lib/systemd/system/daemontools.service << __EOF__
[Unit]
Description=Daemontools service supervision

[Service]
ExecStart=/command/svscanboot /etc/service/
Restart=always

[Install]
WantedBy=multi-user.target
__EOF__

systemctl enable daemontools.service
systemctl status daemontools

qmail.org  is offline, change:

wget http://www.qmail.org/netqmail-1.06.tar.gz

to

wget https://notes.sagredo.eu/files/qmail/tar/netqmail-1.06.tar.gz

Hi,

This is the little multilog patch I would like to share.

By default multilog only support max 16MB log file, for my email server, 16MB only can save 5 hours log, so I made this little patch, and change multilog file limit to 100MB, here how to apply:

cd /usr/local/src
wget https://www.phpini.com/files/multilog_filesize_limit.patch
cd /var/qmail/admin/daemontools
patch -p1 < /usr/local/src/multilog_filesize_limit.patch
package/install

and then update log/run file/s, now can assign file size to s104857600 (100MB).

thank you, Sam.. I'll add it asap

I have recently installed qmail from all the great information here (thanks for this).  But after reviewing my log files they are all 16M and rotate out after the 10th one (current log and then 9 of the renamed @.....#.s files).  So I did check the multilog.c file from the patch just to confirm and everything looks good there.  But it is still cutting the logs out at 16M instead of 100M.  How do I go about figuring out where to change that value and maybe even increase the number of logs it keeps?

in the log run file you have something like this

  exec /usr/local/bin/setuidgid $LOGUSER /usr/local/bin/multilog t n5 s16777215 $LOGDIR \

where n5 is the number of log files and s16777215 is the maximum size in bytes. Look here for more info http://cr.yp.to/daemontools/multilog.html

the creation of clear service is non understandeable and not well described

cat > clear/run < 

syntax error near unexpected token `newline'

what must do run to clear ?

no link creation for service clear

thank you, corrected

I tried installing daemontools exactly as outlined above.  But qmailctl start fails because it can't find svc, svok, etc.  I researched on the web, and it appears that the issue is with daemontools not being found in the PATH.  I've tried updating the path in /etc/profile, and in /etc/inittab (when launching svscan).  Although ps -ef | grep svscan demonstrates that svscan is running, qmailctl can't seem to find scripts in the /command folder.

(Per your suggestion, I'm trying to do this install on Slackware  (version 13.37).)

I also tried running ./rts > rts.out (as mentioned in Dave Sill's book, "Life With Qmail") but it bombs on every line--again complaining 'command not found' one line after the next.

What am I doing wrong?

Thanks in advance for your response.

-Dave Martin

Hi Dave, first of all thanks a lot for your support.

You should have this PATH inside the /command/svscanboot script:

PATH=/command:/usr/local/bin:/usr/local/sbin:/bin:/sbin:/usr/bin:/usr/sbin:/usr/X11R6/bin

qmail should start by means of a line inside /etc/inittab

SV:123456:respawn:/command/svscanboot

so, you may want to check that svscanboot was actually launched on boot, or try to run it manually. I prefer to have it in my rc.local

fyi, I have tested everything on a slackware64 13.37

cheers :)

Roberto

if your linux using systemd

remove SV:123456:repawn:/command/svscanboot from /etc/inittab

create a file at /lib/systemd/system/daemontools.service
edit as

[Unit]
 Description=DJB daemontools
 After=syslog.target network.target

[Service]
 ExecStart=/command/svscanboot
 Restart=always

[Install]
 WantedBy=multi-user.target

copy links

ln -s /lib/systemd/system/daemontools.service /etc/systemd/system/multi-user.target.wants

reboot ur system

qmailctl start continues to return errors on line 17: svok: command not found 

However, ps -ef | grep svscan returns the following:

root 1886 1 0 Mar10 ? 00:00:00 /bin/sh /command/svscanboot
root 1948 1886 0 Mar10 ? 00:00:02 svscan /service
root 9379 6093 0 01:12 pts/0 00:00:00 grep svscan

Could there be a permissions issue with the qmail user accounts?  Does the qmail group need to be granted permission to in order to run the daemontool utilities?

It's seems like the daemontools' installation has failed... can you post please:

ls -l /command/
ps axfuww

anyway feel free to contact me in private using the contact button above