Bug 666 - Packaging takes an extremely long time
Summary: Packaging takes an extremely long time
Status: VERIFIED FIXED
Alias: None
Product: Meta-yocto
Classification: Build System, Metadata & Runtime
Component: meta-yocto (show other bugs)
Version: unspecified
Hardware: x86 Multiple
: High normal
Target Milestone: 1.0 M4
Assignee: Richard Purdie
QA Contact:
URL:
Whiteboard:
Depends on:
Blocks:
 
Reported: 2011-01-20 07:55 UTC by Joshua Lock - Disabled
Modified: 2011-07-25 08:27 UTC (History)
3 users (show)

See Also:
OS type for building Yocto: ---
Type of Regression: ---
Verified:
Documentation change: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Joshua Lock - Disabled 2011-01-20 07:55:31 UTC
As pointed out by Gary on the Poky ML, packaging takes an extremely long time:

"Here's what I mean.  Building for my kernel and timing the
various steps, e.g.
   % time bitbake virtual/kernel -c compile

My kernel:
   compile            1m49s
   deploy             0m16s
   populate_sysroot   0m38s
   package           23m20s"
https://lists.yoctoproject.org/pipermail/poky/2011-January/002635.html

This performance impact may have been introduced by recent changes to the package.bbclass to use the rpmdeps command to discover per-file dependencies.
http://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/?id=54e3a82d9c14f3071db4a4f78312c0e31534696c
Comment 1 Joshua Lock - Disabled 2011-01-20 09:19:04 UTC
A second theory is that the bitbake package/server split is causing a lot of context switches for the messages, during package creation - particularly for large packages like the the kernel. See: https://lists.yoctoproject.org/pipermail/poky/2011-January/002638.html
Comment 2 Richard Purdie 2011-02-01 09:57:45 UTC
This was traced to the server end of the worker communication pipe using small buffers. We've increased the size of the pipe and performance improved dramatically.
Comment 3 Joshua Lock - Disabled 2011-07-25 08:27:04 UTC
Verify fixed