I have jobs that are backing up in the queue because one job hangs. All jobs are defined as Async + Detached. They have e-mail checked for the error and late start events. I have no idea where in the job they hang; the script part or the notification part.

Once one job hangs (sits in the queue with a status of 'Running'), all other jobs in that queue sit there. What causes this since they are all set to Async. I thought the purpose of Async was to eliminate any queueing (like jobs in Task Scheduler do).

I hope this re-wording of the problem helps.

Thanks

This is only partially correc. If a sync option is checked the job will run async byt notification actions will not!
This is not exactly the same as it is in the Windows Task Scheduler.

Notification action hanging causes the entire queue to hang.

If you run a script type notification action add LogAddMessage or LogAddMessageEx calls or turn on the tracing option (Tools/Options menu; Log page; "Trace enabled" option) so you can see which particular operation causes the job to hang.
When tracing is enabled you can use [job id].log file in the Performance Data directory to audit job execution and trace.

: I have jobs that are backing up in the queue because one job hangs. All jobs
: are defined as Async + Detached. They have e-mail checked for the error
: and late start events. I have no idea where in the job they hang; the
: script part or the notification part.

: Once one job hangs (sits in the queue with a status of 'Running'), all other
: jobs in that queue sit there. What causes this since they are all set to
: Async. I thought the purpose of Async was to eliminate any queueing (like
: jobs in Task Scheduler do).

: I hope this re-wording of the problem helps.

: Thanks

Sorry for typos the message should read

This is only partially correct. If the "async job" option is checked the job will run
async byt notification actions will not!
This is not exactly the same as it is in the Windows Task Scheduler.

Notification action hanging causes the entire queue to hang.

If you run a script type notification action add LogAddMessage or
LogAddMessageEx calls or turn on the tracing option (Tools/Options menu;
Log page; "Trace enabled" option) so you can see which
particular operation causes the job to hang.
When tracing is enabled you can use [job id].log file in the Performance Data
directory to audit job execution and trace.

: This is only partially correc. If a sync option is checked the job will run
: async byt notification actions will not!
: This is not exactly the same as it is in the Windows Task Scheduler.

: Notification action hanging causes the entire queue to hang.

: If you run a script type notification action add LogAddMessage or
: LogAddMessageEx calls or turn on the tracing option (Tools/Options menu;
: Log page; "Trace enabled" option) so you can see which
: particular operation causes the job to hang.
: When tracing is enabled you can use [job id].log file in the Performance Data
: directory to audit job execution and trace.

No problem with the typos here :)

So, if I understand it correctly the notification portions of an "Async" job are not run async, therefore the job may cause a queue blockage if it hangs during that portion of the job. Since I don't control our SMTP server (or our network), this is something I must consider as being likely to happen.

So if I want a "pure" Async job, with no possibility of queue blockage, I need to get more sophisicated. I can do it by setting up a Parent/Child relationship between the "worker" job and a new "notification" job. I would setup the notification events to "create a semaphore file" in the parent job, and then setup the child job to monitor for the existence of the semaphore file. The child job then interogates the event and job name to determine what notification action is appropriate.

If I am correct (which would be a first), then I will know what to do as soon as you/me figure out the problems with @V"event" and @V"job_name" not working for me.

Thanks for helping here,
Mike

: Sorry for typos the message should read

: This is only partially correct. If the "async job" option is
: checked the job will run
: async byt notification actions will not!
: This is not exactly the same as it is in the Windows Task Scheduler.

: Notification action hanging causes the entire queue to hang.

: If you run a script type notification action add LogAddMessage or
: LogAddMessageEx calls or turn on the tracing option (Tools/Options menu;
: Log page; "Trace enabled" option) so you can see which
: particular operation causes the job to hang.
: When tracing is enabled you can use [job id].log file in the Performance Data
: directory to audit job execution and trace.

Yes, you are correct. Please keep in mind there exist other ways to deal with such issues.

For example you may want to run a job that send email message and it hangs or simply takes a long time to run you want to kill it.
A solution to this would be to run an external program to send the email (Exchange Admin Kit for example includes sendmail.exe utility, there are also many other similar utilities available on the Internet).

RunAndWait command can run such program with a timeout and kill it after the specified time if the program does is still running.

If you don not want to use an external program you can create a "helper" job in 24x7 which you would run as
RunAndWait "24x7.exe /JOB \"helper job id\"", "", timeout, pid

BTW: I would be very surprised if the problem is with the SMTP server. I guess email functions should fail rather than hang if SMTP server does not work properly. What's more likely is a network problem causing the network communication to break while resolving host name or during data transmission.

: No problem with the typos here :)

: So, if I understand it correctly the notification portions of an
: "Async" job are not run async, therefore the job may cause a
: queue blockage if it hangs during that portion of the job. Since I don't
: control our SMTP server (or our network), this is something I must
: consider as being likely to happen.

: So if I want a "pure" Async job, with no possibility of queue
: blockage, I need to get more sophisicated. I can do it by setting up a
: Parent/Child relationship between the "worker" job and a new
: "notification" job. I would setup the notification events to
: "create a semaphore file" in the parent job, and then setup the
: child job to monitor for the existence of the semaphore file. The child
: job then interogates the event and job name to determine what notification
: action is appropriate.

: If I am correct (which would be a first), then I will know what to do as soon
: as you/me figure out the problems with @V"event" and
: @V"job_name" not working for me.

: Thanks for helping here,
: Mike

Greetings,
I'm having the same queue issue. I'm very new to this software, so I'm not familiar with the terms you are using.
My problem is simple: I have 30ish jobs running anywhere from 1 minute to monthly. I have email notification on failure and late start for all of them. They are all asynchronous and the two heaviest ones are detached. The problem is that "something" happens, and every job begins to fail. I don't know how to debug the issue.
I have all my jobs running in the default queue. Is that wrong?
Any help would be appreciated. Thanks.
Ryan

This design is not wrong, but is not very robust either. A better way is to logically separate jobs into job streams or groups and assign each group to a different queue making jobs in one group asynchronous.
You can think of multiple job queues as if you run print jobs using multiple printers. Each printer is independent but still it prints only one job at a time.

Anyway, you need to find out what is "something" that causes jobs to fail or stop running. Start with the Log Viewer and check what events and errors occurred before the problem was detected. Job Queue monitor can also help. If jobs stop running use it to check what you have in the queue and whether the queue keeps up with the job processing.
If you run script type jobs you can turn on job tracing (Tools/Options; Log page) and find out why scripts do not work as expected and where they go wrong.

On the other hand job troubleshooting may be a time consuming process. It may be easier to schedule periodic system restarts or scheduler restarts if that fixes the problem. Since version 3.4.8 24x7 supports such options. Please see Tools/Options menu for more details.

I recommend to everyone to read "Job Processing Flow" topic in the on-line help or printed manual. That topic includes graphical charts explaining how different job types are executed.

: Greetings,
: I'm having the same queue issue. I'm very new to this software, so I'm not
: familiar with the terms you are using.
: My problem is simple: I have 30ish jobs running anywhere from 1 minute to
: monthly. I have email notification on failure and late start for all of
: them. They are all asynchronous and the two heaviest ones are detached.
: The problem is that "something" happens, and every job begins to
: fail. I don't know how to debug the issue.
: I have all my jobs running in the default queue. Is that wrong?
: Any help would be appreciated. Thanks.
: Ryan

SysOp,
Thank you for your quick reply. I did exactly as you said. I took any jobs running daily, weekly, or monthly, and put them into one queue. Then all other jobs (those running more frequently than daily) each have their own queue.
I accidently trimmed the log file (set it to 300 lines rather than 3000) and I lost the period of time where the "something" happened. I will keep my eyes on it, and I will also do the other things you said. Thanks for your help, and I'll be sure to post again if I have further trouble.
Ryan