we also experiencing the same issue, did you find a resolution?

Hi Simon,

You have seen a huge improvement by using additional Process Servers, in your case. As pointed out to you your sizing is a major issue for your particular customer.

Regards,

Renier

Also curious to know if you found a solution. We're on V12p23 with the same issue, although my end-users get it too. Thanks.

We installed V12 p23 last weekend, together with Syracuse 12.8.2.2, MongoDB 4.0.18 & runtime 93.1.17 and, so far, we haven't had the issue since the upgrade.  Of course, it's only 4 days ago so too early to say for sure that the problem has gone away...

What level of Syracuse and MongoDB are you on?

Thanks for the reply. I have Mongo 4.0.18, and for Syracuse, I have 12.8.0 installer, and X3 reports 12.8.0.44-0.

I downloaded the P23 files from the KB page a while back, and didn't notice when we did the update that the runtime and adxadmin had been updated. I applied the updated versions last night which ultimately didn't help. At this moment I can't access the P23 KB article at all to see if I can get 12.8.2.2.

I did end up installing Syracuse 18.2.2 and runtime 93.1.19, but it didn't help. Our issue seemed to be more related to the total number of runtime connections; when it got above a certain number the crashes would start. I added an additional process server as well as adding some memory to the main runtime server, and thusfar the crashes have stopped.

Did you see https://support.na.sage.com/selfservice/viewdocument.do?externalId=105883 "Batch Server Tasks increasingly failing with ECONNRESET errors" ?

There is also possible issue with windows heap sizing, which I'm just about to write up, although would expect interactive users to also have ECONNRESETs in this case

not mandatory if you launch more than 50 batch sessions without interactive ones! :-) It seams that the issue is the  heap sizing.

have you managed to fix this?

Bonjour,

Il s'agit d'un véritable problème connu chez et l'information ne circule pas. Bref, la solution est la suivante, il s'agit d'une clé de registre sur le serveur de traitement qu'il faille actualiser, puis redémarrer le serveur.

Bon courage (et j'attends mon café, car je viens de vous sauver du temps déjà pas mal perdu )

"There is also possible issue with windows heap sizing, which I'm just about to write up" I did write this up and it is documented in the article: Batch Server Tasks fail with ECONNRESET errors.  Sorry for not updating this thread before.    

In fact, if you search the Sage Knowledge base for "ECONNRESET batch" you will find both the articles I mentioned...