Loading...
Loading...
We identified an issue where the clean-up jobs introduced in Washington may cause the unintentional deletion of attachments where, prior to Vancouver, archiving was enabled for a table with attachments and the sys_attachment group of tables was sharded. The issue may impact instances on Washington DC and later releases where: Archiving was enabled for any table with attachments before Vancouver upgrade, and sys_attachment group of tables were sharded, before Vancouver upgrade. The system property glide.attachment.orphan.enable_cleanup is enabled.
The issue cannot be reproduced on demand. This affects instances on Washington, Xanadu and Yokohama (until they are on a fixed version). The instance will be potentially at risk of this issue if the system property (sys_properties) glide.attachment.orphan.enable_cleanup is enabled. This system property, when enabled, initiates the scheduled execution of the two new jobs introduced in Washington DC, as described below: Purge Orphan Attachments (sysauto_script table, sys_id = 305bdd3b776131103b00d86bba5a9954) Delete Marked Attachments (sysauto_script table, sys_id = 79f49cca77f131103b00d86bba5a9988)
This issue has been fixed in Xanadu Patch 7, Yokohama Patch 1, Zurich and later releases. The Purge Orphaned Attachments and Remove Orphaned Attachments jobs are disabled in the fixed release(s). The workaround provided below is also used to mitigate this issue. This workaround involves disabling the property and jobs to no longer be impacted. This can be performed by running the script below: Workaround script (Disables the scheduled jobs and sys_property for the orphan cleanup): The script linked here will disable the scheduled jobs and system property record. This script has to be run in scripts background (/sys.scripts.do) as admin, in global scope. FAQ Q: Can I run this workaround script on my own? A: Yes, it is attached to this KB and admins can run it on their own in scripts background. This should only be run on Washington, Xanadu or Yokohama instances below the above mentioned fixed version(s). It is not recommended to run it on a pre-Washington instance. Q: What does the output look like on a successful run? A: The main area to observe is at_risk = false in the output after running the script. Q: When will the cleanup feature be reenabled? A: This feature will be reenabled in a future release. There is no ETA currently on this.
PRB1849752
Click on a version to see all relevant bugs
ServiceNow Integration
Learn more about where this data comes from
BugZero Plan
Streamline upgrades with automated vendor bug scrubs
BugZero Prevent
Wish you caught this bug sooner? Get proactive today.