No edit summary |
|||
| (6 intermediate revisions by 5 users not shown) | |||
| Line 5: | Line 5: | ||
# Please complete release notes section with the proposed text which would announce and draw attention to this new feature | # Please complete release notes section with the proposed text which would announce and draw attention to this new feature | ||
= Is upstart ready? = | |||
Is upstart ready to handle parallel startup and become a true event-driven system? | Is upstart ready to handle parallel startup and become a true event-driven system? I recall some discussion on that during F10, might want to get Casey Dahlin to weigh in on that. | ||
Harald: upstart-0.5 would be ready for the basic use case of moving essential services from SysVInit style to upstart and start them in parallel. | |||
== QA review - 2009-01-01 == | |||
The Scope section of this feature is somewhat lacking. One of the items that we're looking for is "what would constitute success or failure of this feature", rather than simply a list of packages that must be modified in order to implement. --[[User:Jstanley|Jstanley]] 03:22, 2 January 2009 (UTC) | |||
--- | |||
Harald, have you seen the correponding OpenSUSE efforts ? http://en.opensuse.org/Boottime/Boot_time | |||
There's also the Mandriva effort: http://lists.mandriva.com/cooker/2009-02/msg00131.php | |||
= readahead + SSD ? = | |||
There have been reports that when using SSD drives, readahead only makes the performance worse (because there's 0 seek time with SSD drives). There should probably be an option to disable readahead easily -- or even automatically. | |||
[[User:Ksvist|Ksvist]] 23:52, 24 March 2009 (UTC) | |||
[[User:harald|Harald]]: | |||
Edit /etc/sysconfig/readahead or "rpm -e readahead". Yes, you are right, it would make sense to provide a tool, to easily switch it off. | |||
Latest revision as of 10:11, 25 March 2009
Feature Wrangler Review (2008-12-09)
Please update the following items and change the category of this page to Category:FeatureReadyForWrangler once they are completed. The feature wrangler's have a watch on this page and will see your changes.
- Please complete documentation section. It could be as simple as a few good URLs to upstream's documentation
- Please complete release notes section with the proposed text which would announce and draw attention to this new feature
Is upstart ready?
Is upstart ready to handle parallel startup and become a true event-driven system? I recall some discussion on that during F10, might want to get Casey Dahlin to weigh in on that.
Harald: upstart-0.5 would be ready for the basic use case of moving essential services from SysVInit style to upstart and start them in parallel.
QA review - 2009-01-01
The Scope section of this feature is somewhat lacking. One of the items that we're looking for is "what would constitute success or failure of this feature", rather than simply a list of packages that must be modified in order to implement. --Jstanley 03:22, 2 January 2009 (UTC)
---
Harald, have you seen the correponding OpenSUSE efforts ? http://en.opensuse.org/Boottime/Boot_time
There's also the Mandriva effort: http://lists.mandriva.com/cooker/2009-02/msg00131.php
readahead + SSD ?
There have been reports that when using SSD drives, readahead only makes the performance worse (because there's 0 seek time with SSD drives). There should probably be an option to disable readahead easily -- or even automatically. Ksvist 23:52, 24 March 2009 (UTC)
Harald: Edit /etc/sysconfig/readahead or "rpm -e readahead". Yes, you are right, it would make sense to provide a tool, to easily switch it off.