Notification on Fault Other - Start Impossible

[SamahSohrab607228]

Does anyone know how I can get Automic v9 to send a notification for a JOB that fails with FAULT OTHER - START IMPOSSIBLE?

I get notified at the schedule level when a workflow is unable to start, but not on a job itself.

[Anon Anon]

One way is to use the task’s Properties, Result tab.  Set its OK status to something like ANY_OK and in its Else Execute enter your notification task such as a Script or Notification object as required.  I'm assuming that is is similar to how you are currently being notified of work flow failures from the Schedule.

We do this under V8 so I assume it works the same in more current versions.

[Anon Anon]

A while back, I created a solution to check for any object that ended in FAULT_OTHER in the past 5 mins and to send an email with a list of objects if there are any.

The query in the SQL variables are written for SQL Server, so if you're using Oracle or DB2, they'll need to be rewritten.

Import the .xml and then execute the SCRI.FAULT > Recurring for every 5 mins (Or adjust to a different time frame as needed)

Download Me

(This only works for v9 and later)

Attachment(s)

FAULT_OTHER_Notifications.xml   7 KB 1 version

[don warr]

Is this download still available? I'm getting an error.

[don warr]

Did you ever get a reply on this? I too am looking for a sql query or something to notify of fault_other.

[Dwight Dick]

Really wish the UI could handle displaying faults

[Pete Wirfs]

 

Marks suggestion does not resolve the issue for V9 and later.  A job can fail to launch because of a typo in a variable name, and it doesn't tell anyone. 

 

I have already implemented a solution similar to Jennifer's.  We don't need to run it every 5 minutes though.  We just run it once an hour.

In addition, I put it inside of a workflow and added to it an email alert when an agent goes down.

Pete

[pavan shashikanth guddeti]

Hi Pete,

 

Our Workflows are getting issue with FAULT_OTHER - Start impossible error on Everyday basis and our Operations team are restarting it but can i know how to resolve this issue instead of having manual triggering all the times and how to find out where the exact issue we are facing ?

 

Can you please help me out here.

[SamahSohrab607228]

what is the error cause of the fault_start? when you view details of the failed workflow what's the Last Message.

[laura_albrecht_automic]

Just running into this issue myself and found this very old thread.  Any new solutions on how to address it?  If something doesn't start (FAULT_OTHER Start Impossible) I really don't want to be 15 minutes behind on getting notified about that.  I realize I can change the frequency to a lower interval - but is this the only option at this point?  If so, it's certainly better than nothing I just wanted to check since this thread is 2+ years old. 

TIA.

[Anon Anon]

We also use HP’s Business Service Management (BSM)  product.  While I do not know its full capabilities as it is owned by another group, it checks for blocked workflows and notifications (pages & emails) are sent to the associated support staff.  Perhaps your company has some similar product that could assist with your needs.

[Anon Anon]

Just to add, as I am new to V11 I set up a simple test.  First I attempted to do it the V8 way by adding my notification to the task's Properties Dependencies tab enabling the Else / Execute attributes and it operated the same as V8; it did not execute the notification object.

I next used the task's Postconditions tab with the following and it executed the notification. This seems too easy and obvious so I must be missing something! Obviously this needs more testing but at least it appears to work as needed.

https://us.v-cdn.net/5019921/uploads/editor/ax/r4b5iersir74.png" width="945">

[SamahSohrab607228]

LauraAlbrecht608310  I haven't seen any new solutions to this issue. I use the script that runs a query but that only informs me and not the user who is responsible for the job.
Mark_Hadler_430  That solution works within a process flow and not if a job is being run recurring in a c_period.

If the jobs are running within a process flow and you are worried about the status of all the jobs (this also exists for objects within a schedule object [jsch]). This field exists in the Attributes Tab of the workflow/schedule object. I've used this to send a notification.
https://us.v-cdn.net/5019921/uploads/editor/1a/m34d76nmcm7q.png" width="762">

BTW, when you run a job recurring, as a c_period, the job doesn't Fault Start until the execution, so you aren't notified when you schedule.

[laura_albrecht_automic]

Thanks Mark.  That actually does work, but that means 2 things:

1 - Objects would have to be within workflows or this condition wouldn't be caught.  Another solution would have to be found for that.

2 - All objects would have to have a Post Condition set.  And far as I can tell - there's no way to set that to be automatic like you can in the Settings for the Line Tool and default predecessor status.  This could be painful.  Would definitely have to be something that was checked / validated for cause every single person here is a new user and they'd forget.

The simplest solution seems to be Jennifer's - unless I'm missing something here.

[Anon Anon]

SamahSohrab607228  -  This is not an issue for us since we only allow task activation from Schedule objects and never use recurring in our production Client.

LauraAlbrecht608310 - I agree that the periodic query is probably the "lesser of evils" as this currently exists.  Perhaps this deserves a Product Idea since it's a pretty gaping hole in AE.

1 - For us, this is not an much of an issue for two reasons:

• All of our production tasks must execute within a Workflow and the task in the Schedule has its Result tab use the Else/Execute for a Notification as required.
• This particular failure "never" occurs in production (well okay it perhaps happens once a year if that) for a missing task within a Workflow that is a result of the manual change control process.  It  never occurs for some type of syntax error, i.e. unset variable in a script, as all of our objects are developed by a dedicated group and not by application developers and such. The objects are are fully tested and are not modified in any manner when transported to production.

2 - Editorializing mode on - We tend to stay from any "feature" of AE that cannot be implemented via the various Process tabs so we can use the provided Search dialog and not have to write SQL.  In my view Automic often adds new worthwhile features but either totally neglects or is slow to add the usability portion such setting a default (as you indicated), using some type of Include and supported by the aforementioned Search to perform impact analysis. - Editorializing mode off 

We have however modified a small number of the AE provided object templates to address some of these issues.  Also, at least you can copy condition and action blocks including their sub-blocks and all parameters that include values feature in the Pre/Postconditions to assist with this so you don't have to do extensive drop-down and selection of the needed parameters.

[laura_albrecht_automic]

Jennifer_Jinhong_34 - I'm getting "No Valid AE import XML file has been defined" when I try to import in the FAULT_OTHER_Notifications.xml file.  Should V9 XML files be OK to import into 11.2?  I thought they would be but doesn't seem like it.

[scott_hughes_automic]

Hi Laura,

I don't know if you are still looking at this.  I was looking for a way to query active schedules for tasks that did not start.  Perhaps I will just use the runtime tab, but that requires a mass change of data at the end of a migration :(

Attachment(s)

ticketing.docx   133 KB 1 version

[laura_albrecht_automic]

Yes scott_hughes_automic - I still need to implement a process to capture FAULT_OTHER errors.  The import of the XML didn't work for me. Jennifer_Jinhong_34 - is a new XML file needed for V11.2?  Or if you could cut / paste in the SQL that'd be great.  Thanks.

[scott_hughes_automic]

I am not sure if Jennifer will answer.  She has moved to a different area of the industry at a new employer.

I just use a SQLI to hold the query in my doc and run a script on a period object to detect results.  I have to log the hits into a vara so that I do not ticket on duplicates.  I can get you an export later tonight

[laura_albrecht_automic]

Oh, WOW.  Didn't know that.  Hope she is doing well.  We will miss her for sure!

Yeah, if you have time to do the export, that'd be great.  I don't need the ticket part for now - maybe later - that'd be cool.

[Anon Anon]

meanwhile I can provide you with mine - runs once a day and comes per mail

https://us.v-cdn.net/5019921/uploads/editor/3j/dhsrtqv4r3ch.jpg" width="1265">