<< < 2020 / 8 > >>
Techleads meetings: Mondays 5pm UTC and Thursdays 9:30am UTC
your name here
add your own
Panda porter boards in the DC
PowerPC toolchain bugs
Building Ubuntu ourselves
Hardware pack development and Landing Teams
Actions from previous meeting
- Andy Green - present
- Angus Ainslie - present
- Amit Kucheria - present
- David Rusling - traveling
- James Westby - present
- Jamie Bennett - present
- Michael Hope - present
- Alexander Sack - vacation
- Christian Reis - traveling
- Jesse Barker - present
- Joey Stanford - present
- Lee Jones - absent
- Loïc Minier - present, tardy
- Paul Larson - present
Paul McKenney - absent
- Sachin Gupta - excused
- Scott Bambrough - present
- Stephen Doel - traveling
- Steve Langasek - present
- Quick discussion about the panda boards. They are almost ready but not quite yet. Target is Friday.
- PPC toolchain bugs
- bug found in the toolchain building the kernel for ppc in Ubuntu
- it's way out there in terms of us providing support
- we can't reproduce. we don't have the hardware. we don't want the hardware
- it's not related to arm
- discussion surrounding how to deal with this
- Loic believes that it's low priority to fix this
- MH is thinking that having a plain vanilla Linaro deb pkg to compare against would help
- Building Ubuntu ourselves
- James W thinks that if David M can get his builders up and running then there might be enough capacity there to rebuild the entire archive.
- Loic thinks we need native hardware to do rebuilds
- MH found qmeu 3 times slower than the clock speed
- Andy reported Fedora had the same findings
- This effort could help us find regressions ourselves and thus, before someone else finds them
- James W mentioned the OBS guys might be helpful.
- HWPACK and Landing Teams
- The amount of hoops we need to go through to build one and get it maintained is excessive. Can we streamline it all?
- discussion surrounding how hwpacks are created
- no automatic way to attach hwpacks at the moment
- Loic said that no steps are really hard but some steps are a little trying, e.g. pinging IS for everything
- Angus said that once it's on snapshots it's a lot less painful but getting it on snapshots via IS takes a week (24 to 36 hours via IS but a week for Linaro due to timezones)
- James W said that once our own instance (not the one we share with OEM) is up these issues will for the most part disappear.
- Offspring HW is awaiting setup in London and Joey expects it to up by the 18th.
- James W said there is some offspring coding needed to be able to automate this better for Linaro.
- James W thinks the main problem is in the communications and exceptions between the landing teams and the platform.
- Loic said that we should be escalate after 2 days vs 9 days
Scott & James W to take this offline and find solution
- Jamie mentioned that we have a lot of hwpacks and we should think about consolidating them. Scott replied that George thinks we should have only four.
Actions from this meeting
Scott & James W to talk about hwpacks and offsprint and report back with recommendation on improvements.
EngineeringUnits/Management/Meetings/2011-03-10 (last modified 2011-03-10 18:42:56)