IZ604: sdmadm memory usage too high

[easymf]

[Imported from gridengine issuezilla ​http://gridengine.sunsource.net/issues/show_bug.cgi?id=604]

        Issue #:      604          Platform:     Sun           Reporter: easymf (easymf)
       Component:     hedeby          OS:        All
     Subcomponent:    cli          Version:      1.0u2            CC:    None defined
        Status:       NEW          Priority:     P2
      Resolution:                 Issue type:    ENHANCEMENT
                               Target milestone: 1.0u5next
      Assigned to:    adoerr (adoerr)
      QA Contact:     adoerr
          URL:
       * Summary:     sdmadm memory usage too high
   Status whiteboard:
      Attachments:


     Issue 604 blocks:
   Votes for issue 604:     Vote for this issue


   Opened: Tue Nov 18 06:40:00 -0700 2008 
------------------------


   Description

   each call to "sdmadm" starts a java process with default heap size (64m). this
   brings some negative consequences, esp. for testsuite running on a zones
   (performing
   a "sdmadm ss" on 10 zones on a machine takes 640m!).

   Evaluation

   should not be a problem for deployment, but is real one for tests.

   Suggested Fix/Work Around

   no workaround, has to be changed.
   enhance "sdmadm" script to support lower heap sizes. ideally, the 8m heapsize
   should be used (-Xmx32m -Xms8m). support of bigger heap size can be solved
   in different ways:
   1. providing a special global switch that will take java args "sdmadm -ja
   [-Xmx32m -Xms32m]"
   2. providing a special global switch that will take just number specifying
   heapsize e.g. "sdmadm -hs 32"
   3. providing a special preconfigured global switch that will specify heapsize
   e.g. "sdmadm -32" or "sdmadm -256"

   Analysis

   "sdmadm" scripts executes java process without specifying a heap size, which results
   in default size to be used (64m).

   How to test

   Depenging on a chosen solution - write a testsuite that will test introduced
   global switches, if any.

   ETC: 2 PD
               ------- Additional comments from rhierlmeier Wed Nov 25 07:21:11 -0700 2009 -------
   Milestone changed