On Tue, Jan 07, 2014 at 11:21:41PM +0200, Mantas Mikulėnas wrote: > On Tue, Jan 7, 2014 at 11:19 PM, Warpme <war...@o2.pl> wrote: > > On 06/01/14 21:52, Zbigniew Jędrzejewski-Szmek wrote: > >> > >> journalctl -o verbose for mysqld-check.sh > > > > Zbyszku, > > > > Here is output for 'journalctl SYSLOG_IDENTIFIER=mysqld-check.sh -o verbose' > > > > Sat 2014-01-04 23:32:07.364159 CET > > [s=bd641dc956b04278a26115fa814b910a;i=4365;b=1fad31caaf1e48348c9d4403b7d68ff6;m=2689289;t=4ef2c98c3163f;x=dc153004a > > _TRANSPORT=stdout > > PRIORITY=6 > > SYSLOG_FACILITY=3 > > _UID=0 > > _GID=0 > > _MACHINE_ID=beb0f0a1c2c303b17213bb8d4a87a3a6 > > _HOSTNAME=mythtv > > _CAP_EFFECTIVE=1fffffffff > > _BOOT_ID=1fad31caaf1e48348c9d4403b7d68ff6 > > SYSLOG_IDENTIFIER=mysqld-check.sh > > MESSAGE=--> mysqld: No need to check MythTV DB. Stop was clean... > > _PID=807 > > _COMM=mysqld-check.sh > > > > And here is exemplary output for mysqld: > > > > Mon 2014-01-06 16:24:36.037183 CET > > [s=6d5ef6e5ef9e413fbcb7fca114dd8465;i=43cb;b=5bbb0bb34be64b14886ed38600badb66;m=41e23ff;t=4ef4edb81a63f;x=b9e5fbfb9 > > _TRANSPORT=stdout > > PRIORITY=6 > > SYSLOG_FACILITY=3 > > _UID=0 > > _GID=0 > > _SYSTEMD_SLICE=system.slice > > _MACHINE_ID=beb0f0a1c2c303b17213bb8d4a87a3a6 > > _HOSTNAME=mythtv > > _CAP_EFFECTIVE=1fffffffff > > SYSLOG_IDENTIFIER=su > > _COMM=su > > _EXE=/bin/su > > _CMDLINE=/bin/su mysql -c /usr/bin/mysqld > > --defaults-file=/etc/mysql/my.cnf --datadir=/var/lib/mysql > > --socket=/var/run/mysqld/mysqld.sock > > _SYSTEMD_CGROUP=/system.slice/mysqld.service > > _SYSTEMD_UNIT=mysqld.service > > _BOOT_ID=5bbb0bb34be64b14886ed38600badb66 > > _PID=1478 > > MESSAGE=Version: '5.5.24' socket: '/var/run/mysqld/mysqld.sock' port: > > 3306 Source distribution > > It could be that the script exits before journald has a chance to read > its cgroup information. That's what I'd guess too. Does is get "fixed" if you insert 'sleep 1' in mysql-check.sh before it exits?
This seems likely because the output suggests that the script is short-lived. > > SYSLOG_IDENTIFIER=mysqld-check.sh > > MESSAGE=--> mysqld: No need to check MythTV DB. Stop was clean... > – it's a known limitation which AFAIK needs > minor kernel changes (or maybe kdbus?) The addition of additional information in the style of SO_CREDENTIALS for normal sockets was rejected by Eric Biederman with "I don't like you" as an explanation. So going through kdbus is the next option. Zbyszek _______________________________________________ systemd-devel mailing list systemd-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/systemd-devel
