ubuntu-bugcontrol team mailing list archive

[TJ <ubuntu@xxxxxx>]

Due to other commitments I allowed my membership (2006-2010) of the bugcontrol team to lapse some time ago. I'm now in a position to devote time to bug-control once more and am therefore requesting
membership.

1. I'm always polite and cannot recall ever having a bad experience with a bug reporter. I signed the code of conduct in 2006.

2. I have read and understand the documentation and have applied it extensively to bugs over a long period of time. I've re-read it now to ensure I'm up to date.

3. Sensitive data such as passwords or personally identifying information (name, account numbers, digital certificate keys) can sometimes be found in stack-traces as data in function arguments or in
the binary data of core dumps.

4. I tend to roam the packages tackling bugs that lack love or appear to be particularly challenging to analyse and trace. Sometimes they come to my attention due to affecting me; other times I might
read about them in forums or mailing-lists or see them mentioned on IRC.

5. Bugs I've worked on. As well as triaging I generally go after the cause of the bug and document my research (for others to follow on) even if I can't provide a final bug-fix or work-around. Where I
can provide a fix
I'll publish a patch and debdiff or link a code branch and request an SRU where appropriate.

5.1. HIGH. [Precise] sudo. Abort in libpam-mount due to pam_open_session() not being called.

Analysis and tracing revealed this as a high importance bug because it shows that sudo was not calling pam_open_session() when a valid cached timestamp was present. This could theoretically expose new
vectors for privilege escalation.

https://launchpad.net/bugs/927828

5.2 NORMAL/HIGH. [Lucid] ubuntuone-client. Ubuntuone-client software wont start.

This was an 'annoyance' bug but also very visible to affected users since it prevented access to ubuntu-one. I delved into the Python code and identified a coding problem, published a corrective
workaround, and alerted the Ubuntu One team.

https://launchpad.net/bugs/666608

5.3 HIGH. [Oneiric, Precise] mountall. could not mount /dev/mapper/cryptswap1

This was a high importance bug that prevented encrypted swap being mounted during boot. I was able to figure out a method of tracing the early start-up to aid running and monitoring the processes
manually. That led to revealing one underlying cause (there were two). I was able to develop a fix which was polished by Steve Langasek. It also caused me to write analysis scripts to determine the
most efficient patch. It turned out the bug report covered 2 distinctly different problems which masked each other.

https:/launchpad.net/bugs/874774

5.4 HIGH. [Oneiric] casper, upstart. PXE/NFS boot requires "IPAPPEND 2" in PXE menus

For those affected this is of high or critical importance since it caused a failure to boot of PXE clients. This affected my network so I was the original reporter too. A simple report detailing a
package which broke existing functionality after upgrade without any warning. This would mainly affect network and systems administrators.

https://launchpad.net/bugs/923219

5.5 HIGH. [Feisty onwards] apache2 2.1.5+. TCP_DEFER_ACCEPT causes random HTTP connection failures in load-balanced web-server farms

In a web server-farm scenario that is fronted by hardware load-balancers, in this case Juniper Redline aka DX, where the load-balancers are configured to use TCP multiplexing (holding open and
re-using HTTP connections to the web servers) there exists the potential for random, unexplained and untraceable connection failures.

This bug was high priority for the e-commerce retailer since it affected payment transactions amongst other things. I spent over a week working on it, writing custom C tools and custom kernels, until
I was able to finally identify the cause (weakly written standards and poor implementation) and suggest a fix (a one line Apache config change).

https://launchpad.net/bugs/134274

Any questions please ask.

TJ.