Bug 1010723 - [PERF] oo-accept-node should cache lscgroup output
[PERF] oo-accept-node should cache lscgroup output
Product: OpenShift Online
Classification: Red Hat
Component: Containers (Show other bugs)
Unspecified Unspecified
unspecified Severity unspecified
: ---
: ---
Assigned To: Jhon Honce
libra bugs
Depends On:
  Show dependency treegraph
Reported: 2013-09-22 15:19 EDT by Andy Grimm
Modified: 2016-11-07 22:47 EST (History)
2 users (show)

See Also:
Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of:
Last Closed: 2013-10-17 09:30:28 EDT
Type: Bug
Regression: ---
Mount Type: ---
Documentation: ---
Verified Versions:
Category: ---
oVirt Team: ---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---

Attachments (Terms of Use)

  None (edit)
Description Andy Grimm 2013-09-22 15:19:08 EDT
Description of problem:

On systems with a large number of users, oo-accept-node can take several minutes to run.  The longest operation is check_users, where there is one invocation of lscgroup per user.  It's possible to run lscgroup once and cache the results instead.  In my testing on a node with 2500 users, this cut the runtime of oo-accept-node from 2:30 to 1:15.  On systems with more users, the difference will be much more pronounced.

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

Comment 2 openshift-github-bot 2013-09-25 15:18:16 EDT
Commit pushed to master at https://github.com/openshift/origin-server

Bug 1010723 - Only run lscgroup once for check_users


Running lscgroup once per user gets expensive, so run it once and
cache the results.  This is also more in line with how quota info
and other data is gathered in oo-accept-node.
Comment 3 chunchen 2013-09-26 21:56:25 EDT
It's fixed, verified on devenv_3831, please refer to the following results:

1. Create 800 apps on devenv instance
2. SSH into the instance and run "time oo-accept-node" at first time
3. run " time oo-accept-node" again

at step 2: the runtime is 1:02
at step 3: the runtime is 0:28

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