Closed savaresejt closed 1 month ago
This is a problem with z/OSMF. We created various user settings to control this behavior, such as listBeforeDownload
to avoid z/OSMF logging errors when files could not be found or maximumParallelFileDownloads
to reduce the number of parallel download attempts, as each request might allocate a new address space. We recommend using IBM Remote System Explorer instead of z/OSMF.
See more details in https://ibm.github.io/zopeneditor-about/Docs/interact_zos_zopeneditor.html
Without switching to RSE could we set the listBeforeDownload to true and set maximumParralellDownloads to a low number to avoid this issue?
I hope so. z/OSMF should return those address spaces automatically as well. If that does not happen it might be worse a ticket against the z/OSMF teams.
Btw. I should have mentioned this doc page: https://ibm.github.io/zopeneditor-about/Docs/knownissues.html#using-z-osmf-with-z-open-editor-and-zowe-explorer that also links to this blog post. Let me know if this helps.
Thanks, this helps a lot! We will take a look at these settings and adjust them if needed.
Do you think the latch contention may have resulted from too many address spaces being spun up and contention in that area?
Development environment used
java -version
and paste the details here):Problem Description
Detailed steps for reproducing the problem:
Observed behavior
User reported having issues saving datasets with zowe explorer. We saw in the log thousands of error messages trying to locate datasets in ++include and copybooks. We observed ~13,000 dead address spaces from the tso user. This has happened 3 times in the last month.
OUTPUT FROM D GRS,C,L
These commands recovered us
Expected behavior
We would still like to figure out the root cause here, this could cause issues if we move zopen editor up to production if we're getting system latches.
The zapp file users are using is:
-