logrotate - zmlogswatchctl: line 114: kill: (1054) - No such process

[zmailcc]

Has anyone else seen this? It seems to be somewhat related to some previous postings (e.g. [SOLVED] Daily Message). In my case, logswatch seems to be running but I occasionally receive some odd emails from the cron daemon.

I get a daily email from the cron daemon. Most of the time it just indicates that some services have been stopped and restarted, which looks normal.

/etc/cron.daily/logrotate:

Stopping logswatch...done.
Starting logswatch...done.
Stopping swatch...done.
Starting swatch...done.
Stopping zmmtaconfig...done.
Starting zmmtaconfig...done.

Occasionally the email contains error messages that appear to be due to a loop in zmlogswatchctl attempting to kill processes that are no longer running.

/etc/cron.daily/logrotate:

Stopping logswatch.../opt/zimbra/bin/zmlogswatchctl: line 114: kill: (1054) - No such process
/opt/zimbra/bin/zmlogswatchctl: line 114: kill: (1055) - No such process
/opt/zimbra/bin/zmlogswatchctl: line 114: kill: (1056) - No such process
/opt/zimbra/bin/zmlogswatchctl: line 114: kill: (1057) - No such process
/opt/zimbra/bin/zmlogswatchctl: line 114: kill: (1058) - No such process
/opt/zimbra/bin/zmlogswatchctl: line 114: kill: (1059) - No such process
/opt/zimbra/bin/zmlogswatchctl: line 114: kill: (1056) - No such process
/opt/zimbra/bin/zmlogswatchctl: line 114: kill: (1057) - No such process
/opt/zimbra/bin/zmlogswatchctl: line 114: kill: (1058) - No such process
/opt/zimbra/bin/zmlogswatchctl: line 114: kill: (1059) - No such process
/opt/zimbra/bin/zmlogswatchctl: line 114: kill: (1057) - No such process
/opt/zimbra/bin/zmlogswatchctl: line 114: kill: (1058) - No such process
/opt/zimbra/bin/zmlogswatchctl: line 114: kill: (1059) - No such process
done.
Starting logswatch...done.
Stopping swatch...done.
Starting swatch...done.
Stopping zmmtaconfig...done.
Starting zmmtaconfig...done.

Is this an indication of a problem or just an artifact of the way zmlogswatchctl works? I am not seeing any other indications of a problem.

It appears that the following code segment in zmlogswatchctl builds a list of processes to kill (in the while loop) and some of them disappear before they are killed (in the for loop). The error messages are being generated by the line that reads "kill -15 ${ids[$i]}" (which is line 114).

... code fragment from zmlogswatchctl ...
if [ x"$os" = "xLinux" ]; then
  index=0;quit=0
  ids[$index]="$pid"
  while [ $quit -eq 0 ]; do
    ((index++))
    ids[$index]=$(ps -o pid --ppid ${ids[$index-1]} | grep -v PID | tr \\n ' ')
    if [ ! "${ids[$index]}" ]; then
      ((quit++))
    fi
  done
  for i in $(seq 0 ${#ids[@]}); do
    if [ "${ids[$i]}" ]; then
      kill -15 ${ids[$i]}
    fi
  done
else
  kill -9 $pid
fi