Monitor daemon keeps shutting down.

[Mark Brown]

I don’t know if the two events are related, but after I saved a few changes to the alerts, the monitor daemon for that site would crash. At first it wouln’t stay up at all. I shut down the services and now it stays running for a while (longest was two hours) before crashing again. I’m going to try re-booting the box later tonight, hoping that will fix it.  

But what would cause this?  And what do I do if rebooting doesn’t help?

CL 5.7 on Windows 2003 virtual server.

Topic

[Mark Brown]

This is in the log when the daemon shuts down.  Any ideas what would trigger this?

Process under service ended

Start Command:

E:quovadxqdx5.7integratorbinhcimonitord.exe

Stop Command:

E:quovadxqdx5.7integratorbinhcicmdnt.exe -p hcimonitord -t d -c die

Error Msgs:

[Robert Milfajt]

Did you make a backup copy of your old alert file?

If so, you might trying to restore the backup and see if things stabilize?

If not, you might find it in $HCISITEDIR/versions/Alerts.

Hope this helps,

Robert Milfajt
Northwestern Medicine
Chicago, IL

[Mark Brown]

I might try doing that. I’ve turned in a ticket with Infor. It’d be nice to know what happened.

[Bob Richardson]

Greetings,

Just a stab here:  we had problems when migrating from 5.6 to 5.7 a few years ago with our monitor daemon experiencing problems with alerts

that contained ampersands (run in background) for exec actions.

The engine does that now on its own.

Look for those terminating “&” (ampersands) that direct the monitor daemon to run scripts in the background and remove them/apply and see if that helps here.

In any event R & D ought to be able to shed light on the problem.

You can run the monitor in debug mode to get more information – they have the syntax for you.  Then give them the log – sometimes you can find the problem from the debug log yourself!

Good luck!

[Mark Brown]

Never did find out why, but replace the alert file with a back up cured the problem.  I guess I’ll add new alerts in smaller chunks…

[Jared Parish]

Just thought I’d throw in my 2 cent.  I’ve also had issues with the monitor daemon crashing when I’ve had complicated alerts.  I’ve not found a solution either.

Happening on multiple instances of CL 5.7 Rev3 on Redhat.

- Jared Parish

[Tim Malaney]

As a FYI – we found we had issues when we used the word ALERT (all caps) in any text in an alert. Worked in 5.6 but caused issues in 5.8

[Bob Richardson]

Greetings,

Just to confirm your findings.

The default.alrt file was restructured for 5.8 where the key word = ALERT

signals to the daemon that it has an alert body to load.  We found this out when moving from 5.7R2 to 5.8.5.0.  Newline delimited lines for the alert body – easier to read for a human.

As a rule we do not use the word ALERT anywhere in our alert configurations going forward.

Good luck on your migration!

[Author]

Replies