Outgoing provisioning detection could be made more efficient

Matthew Clark 14 years ago updated by anonymous 9 years ago 10

The outgoing provisioning pending detection should be made more efficient, as it can easily blow out in larger systems. You really only need to check to see if even one record exists as a pending export in an MA - there is an attribute (mms_connectorspace.export_operation) that flags a particular record as needing an export, and this can be queried quite quickly. Worth investigating for next EB version - there may be an even faster way to find exports

Also on this, this query times out after 30 seconds, and no exports will fire if execution exceeds this timeout.

The query execution time can be improved by adding a custom index to the ILM database, but this

Migrated from Help Desk #427

Regarding Help Desk #407:

In Event Broker v2.2.x, pending exports are calculated for all outgoing operation lists, regardless of whether or not the outgoing operation list is enabled or not.

If the above is addressed, this shouldn't be too much of an issue if it is faster to find out all MAs that require an export

Has the mechanism for detecting pending exports to ILM changed?

We'll be reviewing all the changes we've made with you when you return. Keep in mind we haven't made any changes in at least 6 weeks now.

Matt, do you plan to work with Adam to improve the query in v3.0?

This has been replaced in the outgoing pending plugin through the use of a WMI mechanism for checking exports. Marked as resolved as the WMI approach is much more "safe" than making database calls as in previous versions. If performance is found to be an issue, reopen and reconsider approach

Updated remaining estimate

Note that the mechanism had changed upon resolution of the issue. Query was appropriately optimized for the minimum of information.