690601 - Mac OS 10.7 test slave configuration for rev4 minis

Status: RESOLVED FIXED

(Release Engineering :: General, defect, P3)

Description

[John Ford [:jhford] CET/CEST Berlin Time]

We should have Lion (10.7) test slaves to run tests on.  Because I dont know if puppet works on these machines or if the necessary command line/flat file configuration tools exist, I am filing the bug as generically as I can.

This bug is to track the work needed to develop a reproducible set of steps, scripts and/or manifests to configure Macs as 10.7 test slaves.

Comment 1

[Chris Cooper [:coop] (he/him)]

Changed the summary to note that this is explicitly about getting a 10.7 image that will run tests successfully on the rev4 minis. 

We may have rev5 minis in a few months, but those come with 10.7 already installed while the rev4 minis do not.

Comment 2

[John Ford [:jhford] CET/CEST Berlin Time]

I don't know whether we want to test with iLife installed or not, but my understanding is that these machines will come with iLife installed from the factory.  Because the version of iLife that they would ship with is mac app store exclusive, we'd need to use the mac app store to install iLife on future slave reimages, or use a base image that already has ilife installed.

Another option is to measure the performance impact of having iLife installed as well as whether it affects our unit tests.

Comment 4

[John O'Duinn [:joduinn] (please use "needinfo?" flag)]

(In reply to John Ford [:jhford] from comment #2)
> I don't know whether we want to test with iLife installed or not, but my
> understanding is that these machines will come with iLife installed from the
> factory.  Because the version of iLife that they would ship with is mac app
> store exclusive, we'd need to use the mac app store to install iLife on
> future slave reimages, or use a base image that already has ilife installed.

Per meeting w/jhford: we do not want to have iLife installed on these. Therefore we will do the same as we did for the 10.6 rev4 testers - installing OSX from recovery disk so these machines can run in production *without* iLife.

Comment 5

[John Ford [:jhford] CET/CEST Berlin Time]

Attachment: puppet-manifests wip

This is a WIP patch.  I am able to sync against a Lion machine but there are the following issues:

1) can't create python virtualenvs, need to extract the Python SDK from the Xcode 4.1? installer
2) need to figure out what to do with /Users/cltbld/Library/Preferences/com.apple.LaunchServices.plist
3) Figure out why the following files cause puppet errors even though they are present on the puppet file store:
-/Users/cltbld/Library/Preferences/com.apple.dock.plist
-/Users/cltbld/Library/Preferences/com.apple.LaunchServices.plist
-/etc/sudoers
4) need to figure out how to configure some new lion features

Comment 6

[John Ford [:jhford] CET/CEST Berlin Time]

3) was a permissions issue on the puppet master, resolved

Comment 7

[John Ford [:jhford] CET/CEST Berlin Time]

Attachment: script to extract the pkg files that contain Xcode

Looks like the python sdk is in DevSDK/System/Library/Frameworks/Python.framework

Comment 8

[John Ford [:jhford] CET/CEST Berlin Time]

creating python sdk using:

create-dmg.sh Python.framework/ python-sdk-4.1.1 python-sdk /System/Library/Frameworks/

Comment 9

[John Ford [:jhford] CET/CEST Berlin Time]

wget http://pyyaml.org/download/pyyaml/PyYAML-3.10.tar.gz
tar zxf PyYAML-3.10.tar.gz
cd PyYAML-3.10
python2.7 setup.py install --root=root
python2.6 setup.py install --root=root
python2.5 setup.py install --root=root
python2.7
wget http://hg.mozilla.org/build/puppet-manifests/raw-file/aa75d6e91838/create-dmg.sh
sh create-dmg.sh root/Library/Python pyyaml-3.10 pyyaml /Library/
scp pyyaml-3.10.dmg staging-puppet.build.mozilla.org:/N/staging/darwin11-x86_64/test/DMGs/

Comment 10

[John Ford [:jhford] CET/CEST Berlin Time]

(In reply to John Ford [:jhford] from comment #5)
> 1) can't create python virtualenvs, need to extract the Python SDK from the
> Xcode 4.1? installer

resolved, created a python sdk dmg based on Xcode 4.1.1

> 2) need to figure out what to do with
> /Users/cltbld/Library/Preferences/com.apple.LaunchServices.plist

I think this is something to do with a unit test launching iCal.  Anyone know how we created this file in the first place?

Comment 11

[John Ford [:jhford] CET/CEST Berlin Time]

Attachment: puppet manifests

this patch is still a WIP, but this is syncing properly with lion-testing1.

Comment 12

[Zandr Milewski [:zandr]]

(In reply to John O'Duinn [:joduinn] from comment #4)

> Per meeting w/jhford: we do not want to have iLife installed on these.
> Therefore we will do the same as we did for the 10.6 rev4 testers -
> installing OSX from recovery disk so these machines can run in production
> *without* iLife.

This is slightly troubling to me. What is the percieved problem with iLife, and why do we believe that that problem is worth creating a configuration different than every Mac shipped?

Comment 13

[John Ford [:jhford] CET/CEST Berlin Time]

Regarding the Lion application resuming feature, I think we can get around it by cleaning up the following files on each reboot.  I think this is better than disabling the feature altogether, in case we ever implement this resume behaviour.

rm -rf /Users/cltbld/Library/Saved Application State/*

Information from http://osxdaily.com/2011/07/17/delete-specific-application-saved-states-from-mac-os-x-10-7-lion-resume/

Comment 14

[John Ford [:jhford] CET/CEST Berlin Time]

Found a utility for disabling bluetooth on OS X.  Sources from http://www.frederikseiffert.de/blueutil/, rebuilt using

Clients-MacBook-Pro:Source cltbld$ cd 
Clients-MacBook-Pro:~ cltbld$ mkdir blueutil
Clients-MacBook-Pro:~ cltbld$ cd blueutil
Clients-MacBook-Pro:blueutil cltbld$ curl -LO http://www.frederikseiffert.de/blueutil/download/blueutil.dmg
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100 33343  100 33343    0     0  26940      0  0:00:01  0:00:01 --:--:-- 46309
Clients-MacBook-Pro:blueutil cltbld$ open blueutil.dmg
Clients-MacBook-Pro:blueutil cltbld$ cp -r /Volumes/blueutil/Source .
Clients-MacBook-Pro:blueutil cltbld$ cd Source/
Clients-MacBook-Pro:Source cltbld$ xcodebuild -sdk /Developer/SDKs/MacOSX10.6.sdk/ -arch i386 -arch x86_64
scp build/Release/blueutil staging-puppet.build.mozilla.org:/N/staging/darwin11-x86_64/test/usr/local/bin/

Comment 15

[John Ford [:jhford] CET/CEST Berlin Time]

Attachment: add 10.7 to puppet manfiests

This script sets the resolution from within the puppet manifest if its wrong.  I am still deploying the launch agent for times when the machine is not connected to the puppet master

Comment 16

[Chris Cooper [:coop] (he/him)]

Comment on attachment 568781 [details] [diff] [review]
add 10.7 to puppet manfiests

Review of attachment 568781 [details] [diff] [review]:
-----------------------------------------------------------------

Looks good, but should there be an node entry for the lion ref image?

Comment 17

[John Ford [:jhford] CET/CEST Berlin Time]

(In reply to Chris Cooper [:coop] from comment #16)
> Comment on attachment 568781 [details] [diff] [review] [diff] [details] [review]
> add 10.7 to puppet manfiests
> 
> Review of attachment 568781 [details] [diff] [review] [diff] [details] [review]:
> -----------------------------------------------------------------
> 
> Looks good, but should there be an node entry for the lion ref image?

I'll add it and assume that i don't need another review cycle for that :)

Comment 18

[John Ford [:jhford] CET/CEST Berlin Time]

diff -u b/scl-production.pp b/scl-production.pp
--- b/scl-production.pp
+++ b/scl-production.pp
@@ -1553,6 +1553,10 @@
     include talos_osx_rev4
 }
 
+node "talos-r4-lion-ref" inherits "darwin11-x86_64-test" {
+    include talos_osx_rev4
+}
+
 node "talos-r4-lion-001" inherits "darwin11-x86_64-test" {
     include talos_osx_rev4
 }

Comment 19

[John Ford [:jhford] CET/CEST Berlin Time]

This has been deployed.  As we don't have a baseline for this OS, I have nothing to compare to.  All issues with this machine setup should be filed as new bugs blocking bug 693760