Shell Programming and Scripting

I know it is unlikely, but has anyone checked whether or not the server is running at 6am on the day(s) when the cron job doesn't run? If there is some kind of daily maintenance operation that takes the server out of normal multi-user mode during that timeframe, cron jobs that were missed while not in multi-user mode will not be started if the system comes back up to normal multi-user mode after the job's scheduled start time.

Hi Don,

Thanks for your answer.

For now with the help of bukuni and RudiC I've set the cron in this way

Code:

  # crontab -l
0 6 * * * /path/to/scripts/script1.sh > /root/CronLog/cronlog.log 2> /root/CronLog/cronlog.err

The issue is that the cron job has failed every day of the week.

Do you know how to check if the source server was running a maintenance operation or something else when the cron job failed?

The source server model/version is

SunOS Generic_118833-22 sun4u sparc SUNW,Sun-Fire-V245

Thanks

Did you add the set -vx to the script so it prints useful info?
For the server maintenance, you check the server system logs. You could also search for the relevant cron entries there.

You have got Solaris. (BTW a very old Solaris 10, certainly never updated.)
Solaris /usr/bin/date has no -d option. The trailing -d is ignored! (An initial -d would give an error).
Either you have got GNU date installed, then please add the path to it, usually /usr/local/bin/date or /usr/sfw/bin/date or /opt/csw/bin/date
Or install GNU date.
Or try the TZ trick:

Code:

dyear=`TZ=$TZ+24 /usr/bin/date +'%Y'`
dmonth=`TZ=$TZ+24 /usr/bin/date +'%b'`
ddate=`TZ=$TZ+24 /usr/bin/date +%Y-%m-%d`

Hi RudyC. Yes, I've added set -vx. So I'll need to check by tommorow what appears in those logs. Thanks for the help.

--- Post updated at 02:41 PM ---

Thanks for your suggestions, but since I get the date, year, month in destination server (CentOS) where the cron job is defined, then I don't have issues with date command.

The question about SunOS was more oriented how to know, if possible, if it runs some maintenance operation in a timeframe that includes the hour when the cron is configured on CentOS server (04:00 or 06:00 am).

If not I'll wait tommorow if something appears in log.

Thanks again.

I don't remember the name of the log file that would contain the information about when the system changes states, but there has to be one. And, as has already been stated, look for cron's log file as well.

A quick and simple check would be to see what output you get from the command:

Code:

uptime

when run on that server. If it says the system has been up longer than when your last 6am cron job was skipped, that isn't your problem.

Thanks for your answer Don.

On SunOS server where I copy the files from the uptime command gives this output.

Code:

# uptime
  4:28pm  up 839 day(s),  5:42,  1 user,  load average: 0.08, 0.07, 0.07