*Attendees*: Don, Glen, Anton, Riku, Joanna, Matteo, Xinyu, Dave Rodgman,
- Glen: Shared TFC Board
- Chromebooks - returning to get out of customs and then sending back
- Cypress/NXP being integrated in October. Received h/w in lab
- TFC-82: Arthur working on this while have time
- TFC-36 is ready for Arm team to review. Leonardo
- TFC-176: Intermittent failures. Leonardo continuing to isolate this
- TFC-20 integrated including a resolved regression. Any issues noticed?
- Arm provided AMI files. Go back to mbedTLS?
- Matteo: Will we finish TFC-36? Need to make sure this finishes.
- Glen: Agreed
- Joanna: Would like to continue w/ TFC-176 continue at least part
- Glen recommends re-evaluating after his work day today
- Joanna: Agree to keep it as a background task
- TFC-172: In backlog, not any data in the ticket on quantifying slow
or how to replace.
- Xinyu: No longer reproducing. Xinyu to update ticket and we will
- TFC-171: Seems to be solved from other infrastructure improvements
- Joanna: OK to close resolved. Not sure what it was but not
- Arthur may have bandwidth for another task
- Joanna: TFC-87 may be a good one to work on. Currently the team
performs a work-around.
- Moved TFC-87 to SC Approved from Backlog
- *Action: Anton* to evaluate moving TFC-173 to SC Approved.
Attendees: Riku, Don, Glen, Matteo, Shebu, Anton, Joanna, Dave Rodgman
- TFC-20: Git performance - infrastructure changes happened including
Leonardo infrastructure changes. Tested on stage. Should keep an eye on it
for next few days. New machine already added back in. Better performance
will be seen as well since more jobs can focus on builds and not clones.
- Expect scripts: Should wrap up this week
- Joanna: Brought up LAVA timeouts TFC-176. Pass on 2nd or 3rd
attempts. Initial analysis is too many parallel LAVA jobs. Starting out by
increasing timeout. Would prevent re-running jobs.
- Riku: Should add LAVA lab folks to this ticket since adjusting
- Boards status:
- Chromebooks still dealing w/ import issues.
- *ACTION: Don* ask Julius to reject and resend the 3 boards to the
- Cypress and NXP platforms now in the lab
- LAVA team will be updating to latest release with the new board
configs included. Want this done in the next couple of weeks
- Arthur will be available for some other tasks.
- Dave: AMI images almost ready. Expect it soon.
- Glen: Linaro support prepared to copy to AMI's when they are queued
- Glen: Joanna's new list of issues
- LAVA timeout was one.
- TFC-87: Joanna's team reviewing that one. CI reporting ticket. Need
some guidance/access from Leonardo
- Glen will let him know
*Attendees*: Joanna, Xinyu, Matteo, Janos, Glen, Riku, Shebu, Anton, Ben,
- Glen: Follow up on notes
- Glen: Set up sync meeting to hear Riku/Leonardo/Anton/Joanna on
proposing a solution on the git clone performance issue.
- Glen: TFC Kanban board review
- Glen: Chromebooks stuck in customs - working paperwork now
- Glen: Cypress & NXP platforms both underway
- Glen: Performance issues update: (TFC-171, 172, 164)
- Ben: Limited CI Number of jobs to help relieve a performance issue.
- Riku: Impact - slower builds.
- Anton: were we testing on staging?
- Ben: No
- Should we allocate resources to work on performance?
- Joanna: Would work on server scaling versus Expect scripts
- Riku: TF-M build, launches over 100 builds, then git clones turn
into 400 simultaneous git clones - need to re-factor to do clone
- Riku/Leonardo - 1-2 week estimate
- Anton has some ideas - sync w/ him on potential solution. Once
agreed, begin the work.
- Glen: Meeting set up for tomorrow to discuss code coverage state and
how Arm might be able to help.
I'm adding the triage maillist to the thread. As a best practice, let's cc
that list on items like this as it includes the stakeholders that
prioritize OpenCI tasks on a weekly basis(minimum) so it's helpful info in
that decision making.
I see you're already subscribed to the list which is great! :) Reviewing
the Aug 31st sync minutes, Expect scripts were determined to be the
priority. Code Coverage next steps are also discussed. Looks like Glen
was going to set up a sync meeting to further discuss this one... The
minutes could have called out this action more clearly:
- Code Coverage:
- A sync w/ Leonardo and Joanna to discuss next steps/Current status
on CC shall be planned. Glen to set up
> Is it suspended for pure priority reason, or any technical reason?
So with the above said, this is a prioritization decision made by the
triage stakeholders, not technical.
Hope this helps, please let me know if any questions or suggestions on
improving the process. :)
On Thu, 2 Sept 2021 at 06:35, Leonardo Sandoval <
> Hi Sherry,
> In resume, for priority reasons.
> Right now I am working on some pending tickets for TF-A (expect scripts
> migration, TFC-36 <https://linaro.atlassian.net/browse/TFC-36>). Once I
> complete TFC-36 <https://linaro.atlassian.net/browse/TFC-36> and MbedTLS
> work is still on hold, I will move to TFC-7
> <https://linaro.atlassian.net/browse/TFC-7> immediately.
> On Thu, 2 Sept 2021 at 02:15, Sherry Wu <Sherry.Wu(a)arm.com> wrote:
>> Hi Leonardo and Don,
>> Just noticed that https://linaro.atlassian.net/browse/TFC-7 changed to
>> “TODO”. Wondering what’s latest update for the code coverage tool
>> integration on Open CI.
>> IMPORTANT NOTICE: The contents of this email and any attachments are
>> confidential and may also be privileged. If you are not the intended
>> recipient, please notify the sender immediately and do not disclose the
>> contents to any other person, use it for any purpose, or store or copy the
>> information in any medium. Thank you.
Thanks for the escalation.
I see Ben is in the loop, so that's the correct first step. I've also cc'd
the triage maillist to make sure all stakeholders are looped in.
In general, as a best practice, I would suggest having as much quantifiable
data as possible upfront in tickets like this to help better understand the
magnitude as well as how to reproduce. It looks like the tickets have
already started to capture this, but I also see Ben in the ticket
requesting more. Datapoints of interest in my mind:
- Clear details on how to reproduce: The task(s) where noticeable
degradation is seen - is this in parallel to when large builds have kicked
- Tasks invoked and level of degradation: for example, "Gerrit reviews -
previously took xyz seconds/minutes, now taking 20% more time (or 2x, 10x?,
failing & never completing?)," frequency, etc. The more details the
better! This will help determine the priority we place on resolving. :)
Perhaps coming up with a general template for this could be helpful.
- Is there other degradations beyond Gerrit?
Ben is most certainly much more qualified than me in knowing what support
is needed to isolate/resolve, and, as noted in the ticket, he is asking for
more details as well. Let's get the details captured in the ticket(s) in
preparation for next Tuesday's Triage meeting where we can prioritize the
resolution over other TFC tasks. :)
Ben, feel free to chime in and correct any of my assumptions/suggestions.
On Thu, 2 Sept 2021 at 01:41, Xinyu Zhang <Xinyu.Zhang(a)arm.com> wrote:
> Hi Don,
> We found that trustedfirmware.org is getting slow. Daily work of some
> developers would be influenced.
> Could you please help to take a look on this issue? Here is the TFC link:
> IMPORTANT NOTICE: The contents of this email and any attachments are
> confidential and may also be privileged. If you are not the intended
> recipient, please notify the sender immediately and do not disclose the
> contents to any other person, use it for any purpose, or store or copy the
> information in any medium. Thank you.