From Fedora Project Wiki

< FWN‎ | Beats

(create draft for fwn 171)
Line 10: Line 10:
=== Test Days ===
=== Test Days ===


This week's first Test Day<ref>https://fedoraproject.org/wiki/Test_Day:Radeon_2009-04-01</ref> was on the Radeon driver for ATI video cards, which has seen substantial changes for Fedora 11. The turnout was again excellent and many useful bug reports were filed. The developer present was [[DaveAirlie|Dave Airlie]], and [[User:Adamwill|Adam Williamson]] and [[User:Jlaska|James Laska]]  were present for the QA team. The second Test Day<ref>https://fedoraproject.org/wiki/Test_Day:Power_management_2009-04-02</ref> was on power management. There was a good turnout of dedicated Rawhide users, and the results were quite positive: most testers found everything worked and few major bugs were discovered.
This week's Test Day<ref>https://fedoraproject.org/wiki/Test_Day:Radeon_2009-04-09</ref> was on UEFI<ref>https://fedoraproject.org/wiki/Features/EFI</ref>, the modern BIOS replacement technology which sees increased support in Fedora 11. which has seen substantial changes for Fedora 11. This was mostly an 'internal' test day, as hardware support UEFI is not yet widely available.


Next week's Test Day<ref>https://fedoraproject.org/wiki/Test_Day:2009-04-09_UEFI</ref> will be on the UEFI<ref>https://fedoraproject.org/wiki/Features/EFI</ref> support in Fedora 11. UEFI hardware is not yet widely available, so this Test Day is mostly only of interest to Red Hat QA staff with access to testing hardware.
Next week is currently planned to have two Test Days. The first will be a follow up to the earlier Anaconda storage rewrite Test Day<ref>https://fedoraproject.org/wiki/QA/Test_Days/2009-03-05</ref>, to check how progress on Anaconda's storage code rewrite is going and whether the significant issues reported at the earlier Test Day have been addressed. The second will be on the Presto plugin for yum<ref>https://fedoraproject.org/wiki/Features/Presto</ref>, which adds support for deltarpms - incrementally diffed packages. This is planned to be a new feature of Fedora 11 if it works well enough, so please come out to help test it. This feature is of particular interest to those with slower internet connections, or bandwidth restrictions.


<references/>
<references/>
Line 18: Line 18:
=== Weekly meetings ===
=== Weekly meetings ===


The QA group weekly meeting<ref>http://fedoraproject.org/wiki/QA/Meetings</ref> was held on 2009-04-01. The full log is available<ref>http://www.happyassassin.net/extras/fedora-qa-20090401.log</ref>[[User:Jlaska|James Laska]] and [[User:Adamwill|Adam Williamson]] reported no progress this week with the Semantic mediawiki extension.
The QA group weekly meeting<ref>http://fedoraproject.org/wiki/QA/Meetings</ref> was held on 2009-04-08. The full log is available<ref>http://www.happyassassin.net/extras/fedora-qa-20090408.log</ref>[[User:jkeating|Jesse Keating]] reported good progress with the autoqa project: a post-tree-compose monitor has been written, meaning autoqa is now run automatically when a new Rawhide tree is created. [[User:Jlaska|James Laska]] mentioned that autoqa results are currently being sent to a mailing list<ref>https://fedorahosted.org/pipermail/autoqa-results/2009-April/thread.html</ref>. [[User:Wwoods|Will Woods]] mapped out some potential future improvements in the reporting of results. Jesse noted that a major next step will be doing automatic tests following the builds of packages through Koji, to give immediate feedback on issues to maintainers.


[[User:Jlaska|James Laska]] reported that packaging work on the Semantic extension for Mediawiki was progressing, and one package had already passed review. He also noted that he had created a test Mediawiki instance with the extension enabled, but had not yet been able to do much testing. [[User:Adamwill|Adam Williamson]] confirmed that he also had not had time to do much testing.
[[User:Adamwill|Adam Williamson]] reported that, as requested at last week's meeting, he had filed a bug on the issue with creating live USB images for Fedora 11 trees from Fedora 10. It had been closed as a duplicate, and the original bug subsequently marked as closed in Rawhide. The group concurred that it didn't make sense to say an issue that mostly affected Fedora 10 was fixed in Rawhide, so Adam has posted a further comment to the bug asking for clarification.


[[User:Adamwill|Adam Williamson]] noted that the release notes should cover the known problem with generating live USB images from the Fedora 11 Beta live CD images using Fedora 10, caused by the older version of syslinux it contains. The group agreed to file a bug on this issue and add a note to the release notes, linking to the bug report.
[[User:Jlaska|James Laska]] said he had checked in with [[User:Wtogami| Warren Togami]] about the new blog<ref>http://rawhidewatch.wordpress.com/</ref> he had created to monitor issues in Rawhide. Warren said that it is intended to be a low traffic blog that includes imformation on non-obvious (potentially high-impact) problems affecting rawhide users. Submissions can be made to Warren for posting on the blog.


[[User:Wwoods|Will Woods]] and [[User:jkeating|Jesse Keating]] reported steady progress with autoqa work, and hope to have two hooks - post-repo-update and post-tree-compose - running by next week.
[[User:Jlaska|James Laska]] reported that he has not yet been able to discuss the integration of lab-in-a-box and autoqa with [[User:Wwoods|Will Woods]]. He clarified that the idea is to make it possible to do fully automated installation, testing and reporting of test results on Rawhide within a virtualized guest system.


[[User:Wwoods|Will Woods]] brought up the new blog<ref>http://rawhidewatch.wordpress.com/</ref> started by [[User:Wtogami| Warren Togami]] to note major issues in Rawhide, and suggested the QA group should propose to work together with Warren on the project.
The group discussed the state of play with regards to Fedora 11's final release, particularly whether all appropriate bugs are marked as F11Blocker or F11Target. Members of the group are encouraged to help make sure that all sufficiently important issues are marked as blocking one or the other, and non-serious issues are taken off the F11Blocker list so development resources can be directed appropriately to the most important issues. Everyone agreed that a special meeting should be held soon to work on these lists, and it would be best to involve as many members of the QA and Bugzappers communities as possible.


The Bugzappers group weekly meeting<ref>http://fedoraproject.org/wiki/BugZappers/Meetings</ref> was held on 2009-03-31. The full log is available<ref>http://fedoraproject.org/wiki/BugZappers/Meetings/Minutes-2009-Mar-31</ref>. The group agreed that too much time was being spent on discussing moving the meeting time and there were too many problems to declare a move, so they agreed to keep to the present time for now.
[[User:Jlaska|James Laska]] committed to working on a simple system for customization of Test Day Live CD images with useful links and information for the particular Test Day for which they are built.


The group discussed whether the How to Triage page<ref>https://fedoraproject.org/wiki/User:Beland/How_to_Triage</ref> should stay as a single page or be split up. No broad agreement was reached. Eventually they agreed to focus on finalizing the content before worrying too much about organization. Some revisions to the content of the page were agreed upon, and [[User:Beland|Christopher Beland]] made them.
The Bugzappers group weekly meeting<ref>http://fedoraproject.org/wiki/BugZappers/Meetings</ref> was held on 2009-04-07. The full log is available<ref>http://fedoraproject.org/wiki/BugZappers/Meetings/Minutes-2009-Apr-07</ref>. The group thanked [[User:Beland|Christopher Beland]] for his excellent work on the new How to Triage page draft<ref>https://fedoraproject.org/wiki/User:Beland/How_to_Triage</ref>, and discussed further revisions to it. It was agreed that checking for upstream reports when triaging bugs should not be mandatory, but recommended. There was a long discussion on whether triagers should be required to follow up on bugs they set to NEEDINFO state, or whether a separate group of triagers should be responsible for following up NEEDINFO bugs. Eventually it was agreed that front-line triagers should be encouraged to follow up themselves, but a regular sweep could also be made by other Bugzappers to catch bugs in NEEDINFO state that were not being followed up. No agreement was made on whether bugs in NEEDINFO state should be closed after 30 days of inactivity, or 60.


The next QA weekly meeting will be held on 2009-04-08 at 1600 UTC in #fedora-meeting, and the next Bugzappers weekly meeting on 2009-04-07 at 1500 UTC in #fedora-meeting.
Brennan Ashton (comphappy) reported that his Bugzappers metric reporting system is in the process of moving to the Fedora infrastructure, and a beta site should be available in the next few days. He expects the site to be usable by next week, and requests feedback at that time. The group agreed to aim for a 'production' release by May 6th.
 
[[User:Adamwill|Adam Williamson]] reported that he is drafting an email to devel-list to solicit developer input on the use of the priority and severity fields in Bugzilla, and he will send this draft to test-list for the group's approval before sending it to devel-list.
 
The next QA weekly meeting will be held on 2009-04-15 at 1600 UTC in #fedora-meeting, and the next Bugzappers weekly meeting on 2009-04-14 at 1500 UTC in #fedora-meeting.
 
<references/>
 
=== Triage Day ===
 
A successful Triage Day<ref>https://fedoraproject.org/wiki/BugZappers/Triage_days</ref> was held on 2009-04-07 following the weekly Bugzappers meeting. [[User:Adamwill|Adam Williamson]] and [[User:Tk009|Edward Kirk]] helped to train two new triagers, Haase Niels (arxs) and Scott Glaser (Sonar_Guy).
 
<references/>
 
=== DeviceKit testing  ===
 
[[MatthiasClasen|Matthias Clasen]] notified the QA group<ref>http://www.redhat.com/archives/fedora-test-list/2009-April/msg00447.html</ref> that a new version of DeviceKit, which should fix several bugs reported during the previously held Test Day, had landed in Rawhide, and suggested that those who had encountered issues on the Test Day should re-test with the new code.
 
<references/>
 
=== Triage Days on the Wiki ===
 
[[User:Adamwill|Adam Williamson]] apologized for the delay, and announced <ref>https://www.redhat.com/archives/fedora-test-list/2009-March/msg01129.html</ref> that a Triage Day page was now available on the Wiki, explaining the existence and function of the Bugzappers group's weekly Triage Day.


<references/>
<references/>


=== Anonymous test result reporting ===
=== Fedora 11 Blocker bugs review ===


Paul Frields reported<ref>http://www.redhat.com/archives/fedora-test-list/2009-March/msg01225.html</ref> that the infrastucture team had set up a special namespace for future Test Day pages on the Wiki. This namespace allows anonymous editing, so testers will be able to report their results without having to sign up for a FAS account (and so sign the CLA), which some testers were not comfortable with having to do.
Further to the discussion in the meeting, [[User:Adamwill|Adam Williamson]] requested<ref>http://www.redhat.com/archives/fedora-test-list/2009-April/msg00544.html</ref> the group's help in reviewing the list of blocker bugs for Fedora 11 release.


<references/>
<references/>


=== New group members ===
=== Graphics card Test Day metrics ===


Two new members sent introduction emails: Bashir Souid<ref>http://www.redhat.com/archives/fedora-test-list/2009-March/msg01253.html</ref> and Gireesh Sreekantan<ref>http://www.redhat.com/archives/fedora-test-list/2009-March/msg01259.html</ref>.  
[[User:Adamwill|Adam Williamson]] explained<ref>https://www.redhat.com/archives/fedora-test-list/2009-April/msg00549.html</ref> that he had generated metrics for bug reports from the graphics card Test Days, and posted them to his blog<ref>http://www.happyassassin.net/2009/04/08/test-day-metrics/</ref>.


<references/>
<references/>

Revision as of 23:18, 9 April 2009

QualityAssurance

In this section, we cover the activities of the QA team[1].

Contributing Writer: Adam Williamson

Test Days

This week's Test Day[1] was on UEFI[2], the modern BIOS replacement technology which sees increased support in Fedora 11. which has seen substantial changes for Fedora 11. This was mostly an 'internal' test day, as hardware support UEFI is not yet widely available.

Next week is currently planned to have two Test Days. The first will be a follow up to the earlier Anaconda storage rewrite Test Day[3], to check how progress on Anaconda's storage code rewrite is going and whether the significant issues reported at the earlier Test Day have been addressed. The second will be on the Presto plugin for yum[4], which adds support for deltarpms - incrementally diffed packages. This is planned to be a new feature of Fedora 11 if it works well enough, so please come out to help test it. This feature is of particular interest to those with slower internet connections, or bandwidth restrictions.

Weekly meetings

The QA group weekly meeting[1] was held on 2009-04-08. The full log is available[2]Jesse Keating reported good progress with the autoqa project: a post-tree-compose monitor has been written, meaning autoqa is now run automatically when a new Rawhide tree is created. James Laska mentioned that autoqa results are currently being sent to a mailing list[3]. Will Woods mapped out some potential future improvements in the reporting of results. Jesse noted that a major next step will be doing automatic tests following the builds of packages through Koji, to give immediate feedback on issues to maintainers.

Adam Williamson reported that, as requested at last week's meeting, he had filed a bug on the issue with creating live USB images for Fedora 11 trees from Fedora 10. It had been closed as a duplicate, and the original bug subsequently marked as closed in Rawhide. The group concurred that it didn't make sense to say an issue that mostly affected Fedora 10 was fixed in Rawhide, so Adam has posted a further comment to the bug asking for clarification.

James Laska said he had checked in with Warren Togami about the new blog[4] he had created to monitor issues in Rawhide. Warren said that it is intended to be a low traffic blog that includes imformation on non-obvious (potentially high-impact) problems affecting rawhide users. Submissions can be made to Warren for posting on the blog.

James Laska reported that he has not yet been able to discuss the integration of lab-in-a-box and autoqa with Will Woods. He clarified that the idea is to make it possible to do fully automated installation, testing and reporting of test results on Rawhide within a virtualized guest system.

The group discussed the state of play with regards to Fedora 11's final release, particularly whether all appropriate bugs are marked as F11Blocker or F11Target. Members of the group are encouraged to help make sure that all sufficiently important issues are marked as blocking one or the other, and non-serious issues are taken off the F11Blocker list so development resources can be directed appropriately to the most important issues. Everyone agreed that a special meeting should be held soon to work on these lists, and it would be best to involve as many members of the QA and Bugzappers communities as possible.

James Laska committed to working on a simple system for customization of Test Day Live CD images with useful links and information for the particular Test Day for which they are built.

The Bugzappers group weekly meeting[5] was held on 2009-04-07. The full log is available[6]. The group thanked Christopher Beland for his excellent work on the new How to Triage page draft[7], and discussed further revisions to it. It was agreed that checking for upstream reports when triaging bugs should not be mandatory, but recommended. There was a long discussion on whether triagers should be required to follow up on bugs they set to NEEDINFO state, or whether a separate group of triagers should be responsible for following up NEEDINFO bugs. Eventually it was agreed that front-line triagers should be encouraged to follow up themselves, but a regular sweep could also be made by other Bugzappers to catch bugs in NEEDINFO state that were not being followed up. No agreement was made on whether bugs in NEEDINFO state should be closed after 30 days of inactivity, or 60.

Brennan Ashton (comphappy) reported that his Bugzappers metric reporting system is in the process of moving to the Fedora infrastructure, and a beta site should be available in the next few days. He expects the site to be usable by next week, and requests feedback at that time. The group agreed to aim for a 'production' release by May 6th.

Adam Williamson reported that he is drafting an email to devel-list to solicit developer input on the use of the priority and severity fields in Bugzilla, and he will send this draft to test-list for the group's approval before sending it to devel-list.

The next QA weekly meeting will be held on 2009-04-15 at 1600 UTC in #fedora-meeting, and the next Bugzappers weekly meeting on 2009-04-14 at 1500 UTC in #fedora-meeting.

Triage Day

A successful Triage Day[1] was held on 2009-04-07 following the weekly Bugzappers meeting. Adam Williamson and Edward Kirk helped to train two new triagers, Haase Niels (arxs) and Scott Glaser (Sonar_Guy).

DeviceKit testing

Matthias Clasen notified the QA group[1] that a new version of DeviceKit, which should fix several bugs reported during the previously held Test Day, had landed in Rawhide, and suggested that those who had encountered issues on the Test Day should re-test with the new code.

Triage Days on the Wiki

Adam Williamson apologized for the delay, and announced [1] that a Triage Day page was now available on the Wiki, explaining the existence and function of the Bugzappers group's weekly Triage Day.

Fedora 11 Blocker bugs review

Further to the discussion in the meeting, Adam Williamson requested[1] the group's help in reviewing the list of blocker bugs for Fedora 11 release.

Graphics card Test Day metrics

Adam Williamson explained[1] that he had generated metrics for bug reports from the graphics card Test Days, and posted them to his blog[2].