Hi,
I am using a remote agent to execute some jobs. When there is an error in the execution, the error message is not delivered. This happens when the scheduled job runs at its normal time. If I run the job using "run now" and an error is found, the message gets delivered. Any suggestions?

Here are good and bad news about this issue.

Bad news first: This is by design. When you run the job manually it runs in the foreground and all error messages are passed back. When the job runs on schedule it runs in the background and only "Remote job failed. See log file for details" message is delivered. You should lookup schedule.log on the agent computer for more info.

Now good news: This behavior has changed in version 3.4.20, which will be released tomorrow. In that version all remote messages are passed back to the master scheduler.

: Hi,
: I am using a remote agent to execute some jobs. When there is an error in the
: execution, the error message is not delivered. This happens when the
: scheduled job runs at its normal time. If I run the job using "run
: now" and an error is found, the message gets delivered. Any
: suggestions?

The new version still doesn't take the "Action" Send e-mail message when the Event "Job Error" happens. How can I be alerted or notified via email that the job failed?

: Here are good and bad news about this issue.

: Bad news first: This is by design. When you run the job manually it runs in
: the foreground and all error messages are passed back. When the job runs
: on schedule it runs in the background and only "Remote job failed.
: See log file for details" message is delivered. You should lookup
: schedule.log on the agent computer for more info.

: Now good news: This behavior has changed in version 3.4.20, which will be
: released tomorrow. In that version all remote messages are passed back to
: the master scheduler.

The new version still doesn't take the "Action" Send e-mail message when the Event "Job Error" happens. How can I be alerted or notified via email that the job failed?

: Here are good and bad news about this issue.

: Bad news first: This is by design. When you run the job manually it runs in
: the foreground and all error messages are passed back. When the job runs
: on schedule it runs in the background and only "Remote job failed.
: See log file for details" message is delivered. You should lookup
: schedule.log on the agent computer for more info.

: Now good news: This behavior has changed in version 3.4.20, which will be
: released tomorrow. In that version all remote messages are passed back to
: the master scheduler.

The new version still doesn't take the "Action" Send e-mail message when the Event "Job Error" happens. How can I be alerted or notified via email that the job failed?

: Here are good and bad news about this issue.

: Bad news first: This is by design. When you run the job manually it runs in
: the foreground and all error messages are passed back. When the job runs
: on schedule it runs in the background and only "Remote job failed.
: See log file for details" message is delivered. You should lookup
: schedule.log on the agent computer for more info.

: Now good news: This behavior has changed in version 3.4.20, which will be
: released tomorrow. In that version all remote messages are passed back to
: the master scheduler.

It is surely able to send e-mail message (it is available for ages, since version 1.0)

Make sure you correctly configure your email settings. Check the log file for additional email related errors

: The new version still doesn't take the "Action" Send e-mail message
: when the Event "Job Error" happens. How can I be alerted or
: notified via email that the job failed?

I have been used successfully the email message on error for those jobs that are executed in the localhost.
This problem happens when I schedule the job to run in a remote agent. The same job, running in the local host, sends the email. When scheduled to run in the remote agent no mail is sent.
: It is surely able to send e-mail message (it is available for ages, since
: version 1.0)

: Make sure you correctly configure your email settings. Check the log file for
: additional email related errors

What do mean by job error (is this an error in your program, error message printed to some file, etc...)? Do you get an error message in the job log on the scheduler side?

: I have been used successfully the email message on error for those jobs that
: are executed in the localhost.
: This problem happens when I schedule the job to run in a remote agent. The
: same job, running in the local host, sends the email. When scheduled to
: run in the remote agent no mail is sent.

: What do mean by job error (is this an error in your program, error message
: printed to some file, etc...)? Do you get an error message in the job log
: on the scheduler side?

This is the specific example: I go to an FTP site to get a file, if the file is not there using FTPGetFile, an error is logged in the 24x7 log. If I schedule the job to run in the Remote Agent, this is what gets logged:

7/13/2004 10:03:27 0 193 0 Sample Job Agent Broadcast contacted.
7/13/2004 10:03:27 0 193 0 Sample Job Remote job posted for execution.
7/13/2004 10:03:27 0 193 0 Sample Job Connection to Broadcast closed.
7/13/2004 10:03:27 2 193 0 Sample Job Remote job error! Job exit code: -1. An error occurred while executing automation script: Line 2: Download failed for "". The server has returned an extended error: 200 Type set to I 200 PORT command successful 550 test.txt: No such file or directory 5

If I change the host to run in the local host, these are the log messages

7/13/2004 10:16:01 0 193 0 Sample Job Job started.
7/13/2004 10:16:01 0 193 0 Sample Job Error message sent.
7/13/2004 10:16:01 2 193 0 Sample Job Job New Job execution error. Exit code: -1. An error occurred while executing automation script: Line 2: Download failed for "". The server has returned an extended error: 200 Type set to I 200 PORT command successful 550 test.txt: No such file or direc

In the first case, NO MAIL is sent. In the second case, an email is sent.

Excellent, now I see what you are doing.
It says right in the log Connected/Posted/Disconnected.

: 7/13/2004 10:03:27 0 193 0 Sample Job Agent Broadcast contacted.
: 7/13/2004 10:03:27 0 193 0 Sample Job Remote job posted for execution.
: 7/13/2004 10:03:27 0 193 0 Sample Job Connection to Broadcast closed.

In this scenario the scheduler posts the job to the agent and then immediately disconnects. It doesn't care at that point what happens to that job after it is posted.
The solution: Don't use asynchronous job and connection method. Use synchronous instead then in this scenario you will get something like
Connected/Submitted/Job Error/Disconnected/Notification sent

: This is the specific example: I go to an FTP site to get a file, if the file
: is not there using FTPGetFile, an error is logged in the 24x7 log. If I
: schedule the job to run in the Remote Agent, this is what gets logged:
: 7/13/2004 10:03:27 0 193 0 Sample Job Agent Broadcast contacted.
: 7/13/2004 10:03:27 0 193 0 Sample Job Remote job posted for execution.
: 7/13/2004 10:03:27 0 193 0 Sample Job Connection to Broadcast closed.
: 7/13/2004 10:03:27 2 193 0 Sample Job Remote job error! Job exit code: -1. An
: error occurred while executing automation script: Line 2: Download failed
: for "". The server has returned an extended error: 200 Type set
: to I 200 PORT command successful 550 test.txt: No such file or directory 5

: If I change the host to run in the local host, these are the log messages

: 7/13/2004 10:16:01 0 193 0 Sample Job Job started.
: 7/13/2004 10:16:01 0 193 0 Sample Job Error message sent.
: 7/13/2004 10:16:01 2 193 0 Sample Job Job New Job execution error. Exit code:
: -1. An error occurred while executing automation script: Line 2: Download
: failed for "". The server has returned an extended error: 200
: Type set to I 200 PORT command successful 550 test.txt: No such file or
: direc

: In the first case, NO MAIL is sent. In the second case, an email is sent.

: Excellent, now I see what you are doing.
: It says right in the log Connected/Posted/Disconnected.

: In this scenario the scheduler posts the job to the agent and then
: immediately disconnects. It doesn't care at that point what happens to
: that job after it is posted.
: The solution: Don't use asynchronous job and connection method. Use
: synchronous instead then in this scenario you will get something like
: Connected/Submitted/Job Error/Disconnected/Notification sent

Is there a way to get the Job Error in Asynchronous mode. The idea of using a remote host was to centralize all jobs and have under one single screen all our jobs. In this case if we are unable to see the result and get alerted about the status, it doesn't make sense to have the remote host if the master needs to wait for the result.

I think you are trying to reinvent the wheels.

Just schedule all jobs using one master scheduler but assign them to run on remote hosts (as needed). On these hosts run 24x7 in the agent mode (or master mode)
In the main master scheduler options enable HTML status reports and put them in the web server directory (if needed map a network share to the web server or schedule a job to periodically upload files to a remote web server).

If central logging is not enough, in job properties setup automatic notifications (email, page, SNMP, etc….)

This way you can manage, schedule and monitor all jobs in one central place as well as view all consolidated job reports in one place that can be accessed from anywhere (using any web browser).

: Is there a way to get the Job Error in Asynchronous mode. The idea of using a
: remote host was to centralize all jobs and have under one single screen
: all our jobs. In this case if we are unable to see the result and get
: alerted about the status, it doesn't make sense to have the remote host if
: the master needs to wait for the result.

: I think you are trying to reinvent the wheels.

: Just schedule all jobs using one master scheduler but assign them to run on
: remote hosts (as needed). On these hosts run 24x7 in the agent mode (or
: master mode)

I have done that. Precisely that is what triggered this thread of messages. The problem we have is that some processes take 10-15 min in finishing. If the process finish successfuly, then we are done. If there is an error we need to receive an email so somebody can take a look at the process.
If we use the synchronous mode to enable the capability of the job error (as you suggested previously) then the master gets "locked" for whatever amount of time it takes for the job to finish. Here is when we thought on using the asynchronous mode. We liberate the Master scheduler, and transfer the job to a remote host. Doing that, if there is an error NO MAIL GETS DELIVERED.

: If central logging is not enough, in job properties setup automatic
: notifications (email, page, SNMP, etc….)

Central logging is fine, but in the case of asynchronous tasks, it doesn't log if a job error happened in the remote host (refer to earlier messages in the thread).

Again, our need is to be able to have a single scheduler with multiple remote hosts, and whenever an error happens, get an email (no matter if the job was scheduled in sync or async mode).

It gets locked because you probably use single job queue. Configure several job queues and assign jobs for different applications to run in different queues. This way you can also achieve the required concurrency while leaving the rest to the scheduler.

: I have done that. Precisely that is what triggered this thread of messages.
: The problem we have is that some processes take 10-15 min in finishing. If
: the process finish successfuly, then we are done. If there is an error we
: need to receive an email so somebody can take a look at the process.
: If we use the synchronous mode to enable the capability of the job error (as
: you suggested previously) then the master gets "locked" for
: whatever amount of time it takes for the job to finish. Here is when we
: thought on using the asynchronous mode. We liberate the Master scheduler,
: and transfer the job to a remote host. Doing that, if there is an error NO
: MAIL GETS DELIVERED.

: Central logging is fine, but in the case of asynchronous tasks, it doesn't
: log if a job error happened in the remote host (refer to earlier messages
: in the thread).

: Again, our need is to be able to have a single scheduler with multiple remote
: hosts, and whenever an error happens, get an email (no matter if the job
: was scheduled in sync or async mode).