Certification System 201011-6750 boots 6 seconds slower

Bug #737148 reported by Brian Murray
16
This bug affects 3 people
Affects Status Importance Assigned to Milestone
The Ubuntu Boot Speed Project
Triaged
Low
Unassigned

Bug Description

During QA review of bootchart data provided by the Hardware Certification team
it was determined that certification system 201011-6750 boots 6 seconds slower with
a Natty image from 20110314 than with a Natty image from 20110311.

System url:
 https://certification.canonical.com/hardware/201011-6750
Detailed Results:
 http://people.canonical.com/~brian/daily-bootcharts/201011-6750.html

Revision history for this message
Brian Murray (brian-murray) wrote :

The plumbing layer appears to have taken 3 seconds more than the last install and boot.

Revision history for this message
Brian Murray (brian-murray) wrote :

Additionally this system is not running Unity and the desktop time increased nearly 2 seconds.

Revision history for this message
Robbie Williamson (robbiew) wrote :

I noticed /etc/update-motd.d/98-reboot-required was ran and some additional calls to 'sh' and 'sysctl'. Could we get another run to see if this was a single anomaly?

Changed in ubuntu-boot-speed:
status: New → Triaged
importance: Undecided → Low
Revision history for this message
Colin Watson (cjwatson) wrote :

There's no single thing that stands out to me in the plumbing layer; everything just generally seems to be taking that little bit longer. I suspect that differences in individual small processes like sh and sysctl that show up in the bootchart is that they just happen to be over bootchart's threshold for paying attention.

One thing I do notice is that I/O is much spikier during ureadahead. Is the ureadahead profile up to date? You can force this by running:

  rm -f /var/lib/ureadahead/pack /var/lib/ureadahead/*.pack

(*.pack may not exist - don't worry about that) then reboot, leave it up for a minute or so after boot to give it time to prepare the profile, and then run bootchart on the following run.

Revision history for this message
Brian Murray (brian-murray) wrote :

Daniel - could you perform the test that Colin is asking for? Thanks!

Revision history for this message
Daniel Manrique (roadmr) wrote :

Hi all,

Bootchart data is collected weekly with an image from that same week, so sadly we no longer have the troublesome image from 20110314 to test.

We do have a couple of recent runs with an image from 20110321, that could be useful to see if boot times are still slower.

The first submission was run right after installation:

https://certification.canonical.com/hardware/201011-6750/submission/qNfPddoCZG5uQi4

For the second, I updated the ureadahead profile as Colin asked, then rebooted to collect updated bootchart data:

https://certification.canonical.com/hardware/201011-6750/submission/2gJPv3672uidEvS

Bootchart data can be retrieved from the "attachments" link in each submission.

Please let me know if more data is needed or there are problems retrieving this.

Revision history for this message
Brian Murray (brian-murray) wrote :

Looking at the latest results for this system using an image form 20110328.1 we can see that the plumbing layer has decreased in time to 10.03 seconds which is very close to the 9.33 seconds for Maverick.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.