24x7 Scheduler Multi-platform Edition private buid 4.1.252

24x7 Scheduler Multi-platform Edition private buid 4.1.252

Version 4.1.252 private build is now available. This build contains cumulative changes in internal and private builds 4.1.247 - 4.1.252 and is available by invitation only. It will be available for public release as soon as we are done with the quality assurance testing. If you are a registered user and would like to participate in the private build testing and ensuring it works well in your computing environment please send an email to our support account.

Here is a list of cumulative changes:

1. Support for remotely managing non-ODBC database interfaces and profiles.
2. Support for * wildcards in email job triggers.
3. Support for regular expressions in email job triggers. To use regular expressions, specify email subject and/or message filter in the trigger definition starting with @REGEX: prefix, for example @REGEX:[0-9]{3}.+
4. Automatic deletion of read emails in email job triggers.
5. Introduction of Scheduler.raiseError statement in JavaScript jobs. Now you can gracefully fail jobs using user-defined error code and error message.
6. Introduction of @V"err_code" and @V"err_message" macro-variables that can be used in certain job notification actions to obtain error code and error message associated with the last job error.
7. Internal code references to localhost have been replaced to 127.0.0.1 – it has been reported that on some systems localhost is not mapped correctly to 127.0.0.1 which in turn breaks FTP-based user authentication leading to agent's refusal of all remote connection.
8. Several bug fixes in the API used by 24x7 web management console.
9. Fixes for 'too many files open' errors occurring on Solaris systems after running many jobs for a while
10. Fixes for 'too many files open' errors occurring on AIX systems when using real-time HTML job status reports option
11. Several other unspecified bug fixes.

Last edited by SysOp on Wed Apr 02, 2008 11:13 am; edited 1 time in total

(Please tell me if this is the wrong place for bug reports)

In the Email Trigger section:

- Under Polling Interval the manual still states the message is left on the server. The release notes say it's now deleted

- Under message text and subject there's some 'font' control codes which haven't been translated

I don't know if it's meant to be fixed or still on the list, but the HTML Status Report still only creates output when running as the GUI, not the service.

Thanks for noticing that. The manual is out of date and should be fixed. As of 4.1.252 the emails are now deleted, because leaving them on the server creates significant performance issues and if these emails are never deleted, at some point the scheduler begin spending more time for scanning emails than what specified in the email scan interval and takes a long time to find and process new emails.

Another couple of issues with the beta:

- The Queue Monitor flickers the screen but doesn't bring up the job queue window as it used to

- I lost my custom calendar in the upgrade and had to import it again. I thought this was stored in the job database? (The rest of the db seems OK)

I see the calendar now, it's stored in preferences.xml

I followed the upgrade procedure so don't know why I lost the calendar in there. I also lost settings from Tools - Options (e.g. log size and e-mail settings).

Cheers

This is weird--a number of jobs have had the successful exit code changed from 0 to 1. I thought the bug was related to manually disabling the job, but now I can't reproduce it.

The job ran early this morning, returned 0, so was successful.

I ran the job manually, it returned 0, and was marked unsuccessful because the expected exit code had changed.

This has happened to a number of jobs.

Last edited by barefootguru on Tue Mar 11, 2008 5:38 pm; edited 1 time in total

Do you mean the value you see in the settings? You should be able to open the same .DAT file in both old and new versions. Can you check if the job properties look identical?

Sorry, please forget the bug about return codes. I've been through some old copies of the job db and they've been in there a while. Either there's gremlins or I'm losing my mind ;-)

I am able to reproduce the issue with both job and queue monitors on 2 Win XP computers. It works fine on another XP and also on Win 2000 and Linux systems. I haven't' tried other systems yet. Nothing turns up in the trace file that would indicate an exception. I will likely need to send this issue back to development.

Whenever I start the 24x7 service under Windows XP it rotates the log file. I don't know if this is a bug or design decision...?

e.g.

1. service is running and writing to scheduler.log
2. stop service
3. start service. scheduler.log is renamed to scheduler-yyyymmdd-hhmmss.log and a new scheduler.log created

This means viewing the log can entail opening an older copy of the file--e.g. sometimes the 'log' button in the GUI won't display the older entries.

'clear log on startup' in the GUI is not checked.

This is by design. When the current log grows very large, it is renamed and a new one is started. Loading and scanning of large log file takes a long time, which affects scheduler performance.