Opened 12 years ago
Last modified 10 years ago
#565 new defect
IZ2703: if the Execd installation can't start the Execd, the installer should print hints
Reported by: | pollinger | Owned by: | |
---|---|---|---|
Priority: | low | Milestone: | |
Component: | sge | Version: | 6.2 |
Severity: | Keywords: | install | |
Cc: |
Description
[Imported from gridengine issuezilla http://gridengine.sunsource.net/issues/show_bug.cgi?id=2703]
Issue #: 2703 Platform: All Reporter: pollinger (pollinger) Component: gridengine OS: All Subcomponent: install Version: 6.2 CC: None defined Status: NEW Priority: P4 Resolution: Issue type: DEFECT Target milestone: --- Assigned to: dom (dom) QA Contact: dom URL: * Summary: if the Execd installation can't start the Execd, the installer should print hints Status whiteboard: Attachments: Issue 2703 blocks: Votes for issue 2703: Opened: Thu Aug 28 03:18:00 -0700 2008 ------------------------ When the Execd installer detects that the Execd couldn't be started, it should try to find out the error reason and print hints. The error detection should be improved. Currently qping is used to check if the Execd runs, but there seem to be cases where the Execd is accessible by qping for a short time and then fails, e.g. when the QMaster denies the Execd's registration (e.g. because the sent Execd host name and the on QMaster resolves are not equal). A example for an easy to detect error would be: When the execd writes a /tmp/execd_messages.<pid> file, the execd wasn't able to create it's spool directory. The reason could be permission problems or a not valid execd_spool_dir. When the installer can't guess a reason, it should at least print that debug information can be achieved by setting the SGE_DEBUG_LEVEL environment variable.
Note: See
TracTickets for help on using
tickets.