JDE Enterprise One Crash

webdeveloper

Member
JDE Enterprise 9.1 Windows 7 environment SQL Server 2016

?

Our JDE ERP system recently completely went down. Accounting, Finance, HR and all other subsets were unavailable. After much investigation we were told this resulted from either zombies being created and/or memory leaks which could not be located or would be impossible to locate. We are also told there is nothing we can do but hobble along for the next several months until we are able to get 9.2, not until well into 2017. Something seems wrong with this.

My question is, aren't zombies something that are often created and cleaned up regularly, or are at least something that should be looked for and cleaned up to prevent memory overload? An upgrade occurred at the beginning of the year and we were told since then our memory was going up, but not monitored, and also reached beyond 80 percent capacity resulting in memory leaks rendering the system incapacitated for 3 days. Oracle provided a patch to resolve the issue for now, and we are told by our 'tech' specialists we will have to hobble along for the next few months until 9.2 comes out. We can keep 'putting patches up' to keep us functioning to then. In the meantime there are no guarantees the system won't crash again and we won't be crippled.

I'm thinking this isn't right. Someone isn't providing us with all of the right information. And, the system will crash again. If the memory overloaded, or has been getting increasingly high since the upgrade at the beginning of the year perhaps zombies have been created since then and no one was monitoring? I am not especially tech savvy, so if someone can explain this to me in a way that makes sense, I would really appreciate it. And, I may be completely ignorant, and this may be right. If so, I am grateful for a confirmation and appreciate your time. I am simply working on contingency planning for my department should this fail again, because there are people that are under an extreme amount of stress right now resulting from this and I want to help them.
 
One option is:

Use the BSFN readINI (B980055) to read value from the JDE.INI file.

First try:

szSectionName = 'NETWORK QUEUE SETTINGS'
szVariableName = 'OutputDirectory'
szValueForEntryDefault = <result>
if <result> is blank or null

then

szSectionName = 'INSTALL'
szVariableName = 'B9'
szValueForEntryDefault = <result>

append 'PrintQueue" to the right of <result> with the appropriate folder/directory delimiter (generally '/' or '\')

Post Script: Oops - posted to wrong thread
 
Last edited:
You post doesn't make a lot of sense to be honest.

Firstly, SQL Server 2016 isn't certified for any release of E1 so I'm not sure why that would be in production. Secondly, 9.2 was released in October so when you refer to not being able to get it I'm assuming that means that the business will not upgrade until then. Irrespective of that, the 9.2.0.3 tools are available for you now and this is unlikely to be an apps issue the way that you have described it. Finally, in our experience 9.1 on either the 9.1.4.x or 9.1.5.x tools is very stable and we have some very big clients that do not have any significant issues with zombies and system outages.

The short answer is you need to get a second opinion because I can pretty much guarantee that an upgrade is not the solution to your problems. You should be able to resolve your issues with what you have now.
 
Hi,

There's not enough detail here to be specific but along the lines of what Russell has said - your issues are probably not related to your current release level of JDE. The SQL Server isn't a 'certified' release yet but we also implemented SQL 2014 before it was certified and it's been running fine (although we have experienced some locking issues, which have been possibly been resolved by a service pack).

You mentioned an upgrade happened at the beginning of the year - what kind of upgrade e.g. Infrastructure/JDE release/Tools upgrade?

Zombies can occur for many different reasons (code, data, tuning etc.) but if the system is being administered, these can be cleared easily. In the worst case scenario, an administrator would restart the service(s). It would be interesting to know why the system was affected for 3 days? Are you saying the services wouldn't start? If that's the case, some logs would be useful to see.

Thanks/Regards,

Sanj

E9.1, TR9.1.5.5 (Live) and TR9.2.0.3 (in Test), SQL Server 2014, Websphere 8.5, VMs
 
Very surprised to see that system was down for 3 days. Way you described it looks like system is not monitored or you don't have CNC actively monitoring the system. I would analyze the logs when system went down and find possible cause of failure.

Chan
 
Sounds to me like they're trying to run on Windows 7 - a desktop OS - rather than a certified Windows Server version ... ;-)
 
I'm thinking you're not providing OR receiving all the right information either. First of all, Zombies are just processes that are marked as killed. Nobody should be forcing you to 9.2 if you're on a supported version of 9.1 - secondly, SQL 2016 isn't supported on any release. But thats probably a typo. Lastly, any reboot should clear ALL your "memory issues" !!!! but you shouldn't need to do that ! Let us know if theres any more platform information you can provide to point you in the right direction...
 
Back
Top