Bug 1638314 - Profiling job should be a separate post import job.
Summary: Profiling job should be a separate post import job.
Keywords:
Status: CLOSED NOTABUG
Alias: None
Product: Red Hat Gluster Storage
Classification: Red Hat Storage
Component: web-admin-tendrl-commons
Version: rhgs-3.4
Hardware: Unspecified
OS: Unspecified
unspecified
unspecified
Target Milestone: ---
: ---
Assignee: Anmol Sachan
QA Contact: sds-qe-bugs
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2018-10-11 10:34 UTC by Anmol Sachan
Modified: 2018-11-20 20:03 UTC (History)
5 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2018-11-20 20:03:44 UTC
Embargoed:


Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
Github Tendrl commons issues 1062 0 None None None 2018-10-11 10:36:17 UTC
Red Hat Bugzilla 1637977 0 unspecified CLOSED Remove blivet dependency from gluster-integration module 2021-02-22 00:41:40 UTC

Internal Links: 1637977

Description Anmol Sachan 2018-10-11 10:34:55 UTC
Description of problem: Currently profiling Enable/Disable happens as a part of import flow. It is run during tendrl-gluster-integration sds_sync. Time taken by profiling gets added to the import time which increases the overall import time. If profiling job can be run as a separate post import job, the import timing can be reduced.


Version-Release number of selected component (if applicable):


How reproducible: 100%


Steps to Reproduce: N/A 


Actual results: Profiling job is done as par of import process.


Expected results: Profiling job should be a separate job run by the import flow when it finishes and the user should be able to track the progress of profiling job like all the other jobs.


Additional info: This enhancement is towards reducing overall import time for a cluster in WA

Comment 2 Martin Bukatovic 2018-10-24 14:09:24 UTC
Asking for validation of QE understanding of this rfe/bugfix bug:

* Import job doesn't change profiling on gluster side at all.

* User can start import, asking for profiling to be enabled/disabled in the same
  way,   as it's implemented right now, so there are no changes in UI of import
  screen.

* When Import task finishes, the Enable/Disable Profiling job is started if
  needed, so that:
   - start of the enable/disable profiling job is handled by backend (it should
     work even when one closes the browser after starting import)
   - user is notified about the fact, that the profiling job has been
     started, if it's logged in web ui (in the same way as done when user
     uses "enable/disable profiling" switch in cluster or volume list page)

* Cluster is visible in Grafana after import job is finished, with data which
  doesn't require profiling to be enabled.

Is our understanding correct?

Additional question: how to use import cluster API after this change. Complete
example would help.

Comment 4 Shubhendu Tripathi 2018-11-19 06:27:48 UTC
Is this still a valid requirement?

Comment 5 Anmol Sachan 2018-11-19 14:14:53 UTC
@shubhendu, This not a valid requirement anymore. We are keeping the profiling as a part of the import job. Hence, the bz can be closed.

Comment 6 Shubhendu Tripathi 2018-11-20 12:24:05 UTC
@anmol, I would request this to be closed with proper comments.

Comment 7 Anmol Sachan 2018-11-20 20:03:44 UTC
When this BZ was filed, there was plan to shift the profiling out of the import job. But during the recent testing, it was figured out that this wouldn't make much improvement to reduce the import time for WA, which was the sole purpose of bringing in this change. Thus this BZ is no longer required and hence I am closing this BZ.


Note You need to log in before you can comment on or make changes to this bug.