opencompute-developers team mailing list archive

[Jeff Marcom <jeff.marcom@xxxxxxxxxxxxx>]

Jeff Marcom has proposed merging lp:~jeffmarcom/opencompute/update-memory-tests into lp:opencompute/checkbox.

Requested reviews:
  Open Compute Developers (opencompute-developers)

For more details, see:
https://code.launchpad.net/~jeffmarcom/opencompute/update-memory-tests/+merge/187124

This adds a few jobs in memory.txt.in to run the google stress app test and check the mcelog for memory errors.

The mcelog error check is independent of the memory stress test, in that it's possible the google app stress does not catch all errors.

The stress app even suggests looking at the log to be sure.

Also,

stressapptest and mcelog are added as package dependencies.

Tests:
* m: memory/info
* e: memory/mcelog_check
* o: memory/stress_30min
  +: Combine with character above to expand node
   : Space when finished

Output upon introducing an error in mcelog:

   memory/info [pass]
   memory/mcelog_check [fail]
   memory/stress_30min [pass]

Output when mcelog is clean:

Memory tests

   memory/info [pass]
   memory/mcelog_check [pass]
   memory/stress_30min [pass]
 : Space when finished


-- 
https://code.launchpad.net/~jeffmarcom/opencompute/update-memory-tests/+merge/187124
Your team Open Compute Developers is requested to review the proposed merge of lp:~jeffmarcom/opencompute/update-memory-tests into lp:opencompute/checkbox.

=== modified file 'data/whitelists/opencompute-ready-local.whitelist'
--- data/whitelists/opencompute-ready-local.whitelist	2013-08-06 16:41:25 +0000
+++ data/whitelists/opencompute-ready-local.whitelist	2013-09-24 00:13:59 +0000
@@ -29,6 +29,7 @@
 dmidecode_attachment
 efi_attachment
 lspci_attachment
+mcelog_attachment
 meminfo_attachment
 modprobe_attachment
 modules_attachment
@@ -62,6 +63,8 @@
 disk/smart_.*
 __memory__
 memory/info
+memory/stress_30min
+memory/mcelog_check
 __miscellanea__
 miscellanea/ipmi_test
 __networking__

=== modified file 'debian/changelog'
--- debian/changelog	2013-09-13 17:07:37 +0000
+++ debian/changelog	2013-09-24 00:13:59 +0000
@@ -2,9 +2,10 @@
 checkbox (1.16.7~OCP) UNRELEASED; urgency=low
 
   [ Jeff Marcom ]
+  * Added googlestress app memory tests
   * Updated plainbox based on version 0.4.dev in lp:checkbox (16.12)
 
- -- Jeff Marcom <jeff.marcom@xxxxxxxxxxxxx>  Fri, 13 Sept 2013 10:13:04 -0400
+ -- Jeff Marcom <jeff.marcom@xxxxxxxxxxxxx>  Mon, 23 Sept 2013 10:13:04 -0400
 
 
 checkbox (1.16.6~OCP) UNRELEASED; urgency=low

=== modified file 'debian/control'
--- debian/control	2013-08-21 16:16:17 +0000
+++ debian/control	2013-09-24 00:13:59 +0000
@@ -55,7 +55,9 @@
           glmark2,
           glmark2-es2,
           hdparm,
+          libaio1,
           lshw,
+          mcelog,
           mesa-utils,
           nmap,
           obexd-client,
@@ -64,6 +66,7 @@
           smartmontools,
           sox,
           stress,
+          stressapptest,
           sysstat,
           wmctrl,
           wodim

=== modified file 'jobs/memory.txt.in'
--- jobs/memory.txt.in	2012-10-09 15:21:35 +0000
+++ jobs/memory.txt.in	2013-09-24 00:13:59 +0000
@@ -7,8 +7,22 @@
  the size of the memory modules detected by DMI.
 
 plugin: shell
-name: memory/check
-user: root
-command: memory_test
-_description:
- Test and exercise memory.
+name: memory/stress_30min
+user: root
+command: stressapptest -s 1800
+_description:
+ Test and exercise memory.
+
+plugin: shell
+name: memory/stress_1hr
+user: root
+command: stressapptest -s 3600
+_description:
+ Test and exercise memory.
+
+plugin: shell
+name: memory/mcelog_check
+user: root
+command: cat /var/log/mcelog; [ ! -s /var/log/mcelog ]
+_description:
+ Grabs mcelog containing memory errors.