.NET Version on Servers

@Bart_Elia, @aidacra

Thanks for all your support. It has been really frustrating for us.

  • Support has webex’ed numerous times and has not been able to figure it out.
  • We get the issue with all BPM’s and customizations turned off.
  • We have upgraded to .33 and the issue still occurs. Even with BPM’s and customizations turned off.
  • The same jobs always fail in our Live environment and even in our pilot environment after we refresh with Live.
  • Today Support did a webex and ran performance diagnostics on our Epicor servers. They didn’t find anything alarming at first glance. They are still reviewing the results.

We have not tried the other items on your (@aidacra) list because we don’t know how. We take it one step at a time keep asking/calling to inquire the next step.

On top of everything, not every support person updates the case ticket with what they did nor do they call us back after they say they are going to. That makes working with support extremely difficult.

@Bart_Elia, @aidacra

We don’t understand it but found what was causing the issue. We have a simple global alert that was recently created by our customer service. It was to create a memo and send alert when “Operation has been completed”. It had a from email address and a label, “51 Completed”. It was set to go to a email with text that red “Please Flip Quote”.
When this alert is active, our backflush hangs on select jobs. If we deactivate, backflush runs with no issues.

We are working with support with this issue. The reason this was not identified on the Epicor system was because the support person noticed the process hung when they first ran it for us on our uploaded db. They communicated to the tech team and the tech team did something, and then from that point on support had no issue running backflush with or without the alert on. Support was able to identify that they too can get the backflush process to hang and is fixed somehow when she tells the tech team to restart the task agent because a process hung. The support person feels the tech team does something more than just restart the task agent but it has not been communicated.

Still investigating.

Very odd. This was a BPM?

no bpm. Native global alerts.

@Bart_Elia, @aidacra

Well we pretty much got the "sorry, we can’t help you’ from Epicor.

Epicor is able to replicate the issue using our DB however they are not able to in their environment. We were told by support that since development was not able to replicate the issue in their environment that it will be unlikely that we will get a root cause or solution.

I am doing my best not to start venting but this is a prime example of what happens to us time and time again. Now we are unable to use another global alert. We already are unable to use the global alert when a task has been completed (Case CS0000573759 that turned into PRB0190899 which they rejected a fix for 10.1.500.x) and now the global alert for when an operation has been completed.

The support on the phone apologized and said there is nothing much we can do. We will not get an answer as to the what and why to our problem. She said we will just have to stop using global alerts.

…(sigh)… deadend. :cry:

Hi Nathan,

I am particularly interested in the disabling of BPM and removing the form customization. Is this script available for us to use as well?

If yes, it will be a great help to us and probably other customers to reduce the load on Epicor support and do more due diligence on our own before escalation.

Right now manually to consistently do all of the above you mentioned is tricky.

The challenge I have is that anything I publicly put into the world seems to automatically make it something I have to personally support until the end of times–doesn’t matter the disclaimers I put on it or the caveats included that explain only do this if that else don’t ever to this.

Here is a dramatized reenactment of a situation I found myself in not so long ago:

@aidacra posted something eight years ago for a product that hasn’t been updated for seven years, and I just ran against a database for a product that was just released today and something bad happened to my database, @aidacra has to fix it! No, I didn’t read the obvious disclaimer at the top! Or, the “don’t do this without testing it” in the middle of the comment. No, I didn’t backup my database first–why would I have backed up my database first if @aidacra said it was ok eight years ago for an earlier product?!..” --Anonymous

Needless to say, I’m always a little hesitant to publish certain things because I’ve been more than once bitten twice shy by things that would be fine for 99.9% of situations without any modification, but, it’s the 1 out of 1000 where the information provided can actually cause harm unless it is fully understood that I worry about because I don’t want to ever negatively impact anyone by something I say or post–even if I really try to warn people about the 1 in 1000 situation.

That being said, after seeking sage guidance from my resident expert and housemate, if you all pinky swear promise to the following:

  1. don’t ever attempt anything that some random internet person tells you unless you have a 100% fool-proof way of reverting it (aka: don’t do stuff in production unless you are completely sure of what it is ACTUALLY going to do by testing it first, and never do anything permanent against a database without backing up your database beforehand).
  2. that any *past, present or future kerfluffles that potentially could maybe happen because of what I’m about to post is the end users’ responsibility. (aka: it is provided as-is with no support whatsoever for what comes from executing it).
  3. you really accept the fact that Support cannot help you with anything that the script does or doesn’t do.

*time travel could be a thing in the future, and I want to be covered

below is a link to the majority of the tailor db script with some of my sanitized internal comments as to what that statement is generically doing when we execute a script like this against a customer database when we bring it in-house. You will also find a really long disclaimer and the script as provided when executed in its entirety will not actually do anything against a database. For those with experience with TSQL the changes needed to make it work should be straightforward–if you can’t identify the changes needed, then you really shouldn’t be using this script and is best to just not(truly sorry about that, but, it’s really for your own good). I would greatly appreciate no one publicly stating what the changes needed to make the script work in its entirety or republishing this script (or even a subset of the script) on this site, or elsewhere.

Link to the partial tailor db script is here.

Please community, don’t break my heart and take this something I put out there for the betterment of those that can safely use it and hurt yourself.

3 Likes

Don’t MESS with the LAWYERS…

1 Like

Just for the record I get a similar Issue when using “Generate Only” on IEx AP ACH Payments. Everything else works besides the Task Agent creating the ACH.txt file “The underlying provider failed on EnlistTransaction.”. NO DTC Enabled. Everything else works. ACH Works when I do it on Client vs Generating it on Agent.

Just thought id share =) might be related.

I only have the out-of-the-box Basic Local DTC just like on a Fresh Windows install

Should RPT Servers be Network DTC instead of just Local?

Mine’s just Default as-is out of box Windows.

It could all be related that Task Agent does have Max Concurrent Tasks set.

I recall someone else having the ACH Issue it was mainly related with the Output Filename… Anyways; not complaining. Just sharing - it might lead to some useful info.