Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 155484 - PORTAGE_NICENESS not honored during global update stage.
Summary: PORTAGE_NICENESS not honored during global update stage.
Status: RESOLVED FIXED
Alias: None
Product: Portage Development
Classification: Unclassified
Component: Core - Interface (emerge) (show other bugs)
Hardware: All Linux
: High normal (vote)
Assignee: Portage team
URL:
Whiteboard:
Keywords: InVCS
Depends on:
Blocks: 147007
  Show dependency tree
 
Reported: 2006-11-17 06:46 UTC by Duncan
Modified: 2006-11-17 20:54 UTC (History)
0 users

See Also:
Package list:
Runtime testing required: ---


Attachments
emerge --info (emerge.info,3.61 KB, text/plain)
2006-11-17 06:54 UTC, Duncan
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Duncan 2006-11-17 06:46:41 UTC
portage-2.1.2-rc1-r7

Portage doesn't appear to honor PORTAGE_NICENESS during the "Performing Global Updates: <$PORTDIR>/profiles/updates/<quarter>" stage.  I have PORTAGE_NICENESS set to 19, and ksysguard monitoring sys/app/nice CPU usage.  While I normally see a mass of yellow (niced cycles) activity while portage is running, it goes nearly solid blue (application cycles) during the global update stage.  Note that newer portage update speed has improved to the point where it was actually practical to add FEATURES=fixpackages, as I have -- it's just not nicing it in accordance with the settings in make.conf, as it does with pretty much everything else it does.

What prompted me to look up and see what was happening was my Internet radio stream suddenly went jerky.  That doesn't normally happen, even with high (double digit) load averages, when portage is running niced, but whatever it launches to do the updates apparently doesn't honor the niceness settings, as demonstrated by the immediate leap in un-niced process activity when the updates run.

I /think/ the same issue may exist when syncing, as I noticed something a bit odd during the sync as well, but I can't be sure on that.  I'm quite sure on the global updates, as the timing of both the start and end of the spike in "application" activity coincided with the start and end of the global updates stage.
Comment 1 Duncan 2006-11-17 06:54:23 UTC
Created attachment 102205 [details]
emerge --info

Umm... make that -r6, I had just synced and noticed the issue when I did my first emerge, to -r7.  Here's emerge --info
Comment 2 Zac Medico gentoo-dev 2006-11-17 07:03:46 UTC
Thanks, this is fixed in svn r5071.
Comment 3 Zac Medico gentoo-dev 2006-11-17 20:54:21 UTC
This has been released in 2.1.2_rc2.