-
Title
The command fixusers errors out with " shareplex is up, cannot continue" (though Shareplex is sh -
Description
When trying to get rid of redundant Post processes that arise when toggling between single-threaded Post (STP) and multi-threaded Post (MTP), one tries the command fixusers but it results in the following error:
qview> fixusers queue_name+P+o.source_SID-o.target_SID
SharePlex is up, cannot continue
Shutdown SharePlex and try again
qview >
This is despite the fact that Shareplex is not running, nor are there any orphaned Shareplex or Post processes observed on the server for this instance of Shareplex. -
Cause
This behavior in qview typically shows up when running into hung semaphores/shared memory segments problem.
-
Resolution
The following steps need to be carried out:
1. One needs to identify and remove hung semaphores or Shared memory segments before attempting to run fixusers to get rid of redundant Post processes (See Solution # SOL6531 for details).
2. Run fixusers as below and it should succeed:
qview> fixusers eq queue_name+P+o.source_SID-o.target_SID
Remove 2:+PP+queue_name+sp_opst_mt+ o.source_SID-o.target_SID
3. Restart Shareplex and Post. -
Additional Information
See SOL10279, SOL12279, SOL15994, SOL27653, SOL21186 and SOL15994 for various scenarios resulting in the need to run fixusers (which is what one must have encountered to have to run fixusers in the first place)