Bug#852461: systemd: Creating scopes with systemd-run unreliable

[Michael Gebetsroither]

Source: systemd
Version: 215-17+deb8u5
Severity: normal

Dear Maintainer,

The creation of scopes with systemd-run --scope is unreliable.
It seems to depend on system "load" or some other factors affected
by system "load".
On production systems we've seen an error rate of up to 20%.

It feels the same like the often needed patches for software to reliable
connect to unix sockets under high-load.

It would be nice if systemd-run --scope would retry the dbus/unix-socket
connect a few times as the error is only temporary.

% for i in `seq 50`; do (systemd-run --scope -- sleep 5 2>&1 ) & done 2>&1 |grep -ni failed

# First execution
3:Failed to create message: Input/output error
22:Failed to create message: Input/output error
49:Failed to create message: Input/output error

# Second execution
14:Failed to create message: Input/output error
28:Failed to create message: Input/output error
41:Failed to create message: Input/output error

% systemd-run --scope -- sleep 1; echo $?
Failed to create message: Input/output error
1

ps.: i had to manually corrected the system information because reportbug is not
possible on our production servers.

gebi

-- System Information:
Debian Release: 8.6
  APT prefers stable
  APT policy: (500, 'stable')
Architecture: amd64 (x86_64)

Kernel: Linux 3.16.36-1+deb8u2 (SMP w/4 CPU cores)
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd