Hi everyone,
We are running into an issue that has been rather difficult to pinpoint, and I was hoping that opening up the question to the community might yield some additional insight. Ultimately, our goal is to have one set of DNA executables/sqts on a share that DNA Agents on either side of an active-active configuration - installed on the AE servers - are able to access. However, when we set up for this configuration we experience exitValue7 consistently.
At best, we've been able to determine that this means the issue is with our .lis file (one of several reports which will write to a specific directory). The documentation we have on this states:
Problem with LIS file (create/write).
This doesn't appear to be permissions related, however, as a restart of the job in question will complete this create/write without issue. Our network team has not been able to identify any connectivity issues at all between source and destination over the period of most of these errors. It does not appear to be related to specific jobs (we've seen this on many, many jobs), but it does tend to appear more often on longer running jobs or during heavier load (such as month-end processing). We've built a second server to host the DNA executables in case the issue was the original share/server, but we still experienced exitvalue7 under this configuration.
In order to work around this issue, we've had to have the DNA executables residing locally on the Agent servers (while also writing the .LIS files locally), which means two additional copies of executables (aside from a set for Fileloader that we have to keep as-is, residing on a share). Even if we only have one Agent currently active and using the share/writing to share for the LIS file, however, we still experience exitValue7 (though at a reduced rate). Executing/writing locally doesn't *guarantee* we won't see the issue, either, but it's certainly not frequent when we're in this configuration.
We sometimes have up to hundreds of concurrent users to where the DNA executables reside on the share, so I don't believe that it's "two Agents reaching the same executable at the same time." We're currently testing to see if splitting these executables yields less frequent errors, but after this... I'm not sure what else to check. I was hoping that there may be a setting or two that I could check between the servers, but since it works on restart I don't know what settings might be best to look at first.
Any/all ideas are welcome at this point - thanks for taking a look!
------------------------------
[Sr. Systems Engineer]
[BECU]
------------------------------