Breakpad/Status Meetings/2015-10-21: Difference between revisions
< Breakpad | Status Meetings
Jump to navigation
Jump to search
(Created page with "<small> [[Breakpad/Status_Meetings/{{#time: Y-m-d | {{SUBPAGENAME}} -1 week}}|« previous meeting]] — index – Breakpad/Status_Meeting...") |
|||
(22 intermediate revisions by 3 users not shown) | |||
Line 18: | Line 18: | ||
== Operations Updates == | == Operations Updates == | ||
* idling, waiting for the Vault/Atlas stuff to change how we do configuration | |||
* "more stable test environment" | |||
** should we allow writes from local devs??? | |||
** Terraform maybe to the rescue | |||
** Clone stage DBs | |||
*** we think we need ES *and* PG | |||
** "relatively light lift [to get done]" | |||
** lars: get a flow of prod crashes directed to your own thing important | |||
** Might be hard to build for a specific feature branch (e.g. terraform deploy per PR) | |||
== Project Updates == | == Project Updates == | ||
* In two weeks we switch TCBS and Crashes per User, by default, to be the SuperSearch based ones. | |||
** old and new equivalents will have links back and forth | |||
** mbrandt has already switched his tests to run on the reports | |||
* the great poison pill crash of October | |||
** guarding against this in the future? | |||
*** not infinite re-attempts | |||
*** fallback storage should redirect to another rabbitmq or S3 | |||
**** let's file a bug! | |||
*** lonnen: should be use ulimit to cap stackwalker from bloating? lars: yes, and time timeout (subprocess) | |||
**** let's file a bug! | |||
** symbols was probably the cause of the memory bloat | |||
* crash-analysis | |||
** peterbe: still waiting for graphics to ack the new report | |||
** lars? correlation scripts | |||
*** paying some technical debt | |||
*** should be ready for Mozlando | |||
** peterbe to attack the missing symbols job | |||
=== Deployment Triage === | === Deployment Triage === | ||
* http://mzl.la/NuW9Zi | * http://mzl.la/NuW9Zi | ||
** lets release Thursday morning | |||
=== PR Triage === | === PR Triage === | ||
Line 29: | Line 56: | ||
== other business == | == other business == | ||
* Python 3? [adrian] | |||
** webapp, yeah, maybe | |||
** leeroy, well, probably hard | |||
** configman, far away still | |||
* what is our bus factor for AWS? [adrian] | |||
** low risk because lonnen and rhelmer and phrawtzy are still "around" | |||
* extract out socorro-collector and socorro-processors | |||
** Terraform work, earlier this year, planned for this | |||
** common-socorro library and connecting distinct repos by PyPI alike | |||
** breaking out the collector would be "25 lines of code" --lars | |||
** lars is not convinced :) | |||
== Travel, etc == | == Travel, etc == | ||
* lars - out 11-4 through 11-13 for gettin' hitched. | |||
* jp - 31st oct - 5th Nov Conference, 5th - 11th vacation | |||
== Links == | == Links == |
Latest revision as of 18:55, 21 October 2015
« previous meeting — index – next week » create?
Meeting Info
Breakpad status meetings occur on Wed at 11:00am Pacific Time.
Conference numbers:
Vidyo: Stability 650-903-0800 x92 conf 98200# 800-707-2533 (pin 369) conf 98200#
IRC backchannel: #breakpad
Mountain View: Dancing Baby (3rd floor)
Operations Updates
- idling, waiting for the Vault/Atlas stuff to change how we do configuration
- "more stable test environment"
- should we allow writes from local devs???
- Terraform maybe to the rescue
- Clone stage DBs
- we think we need ES *and* PG
- "relatively light lift [to get done]"
- lars: get a flow of prod crashes directed to your own thing important
- Might be hard to build for a specific feature branch (e.g. terraform deploy per PR)
Project Updates
- In two weeks we switch TCBS and Crashes per User, by default, to be the SuperSearch based ones.
- old and new equivalents will have links back and forth
- mbrandt has already switched his tests to run on the reports
- the great poison pill crash of October
- guarding against this in the future?
- not infinite re-attempts
- fallback storage should redirect to another rabbitmq or S3
- let's file a bug!
- lonnen: should be use ulimit to cap stackwalker from bloating? lars: yes, and time timeout (subprocess)
- let's file a bug!
- symbols was probably the cause of the memory bloat
- guarding against this in the future?
- crash-analysis
- peterbe: still waiting for graphics to ack the new report
- lars? correlation scripts
- paying some technical debt
- should be ready for Mozlando
- peterbe to attack the missing symbols job
Deployment Triage
- http://mzl.la/NuW9Zi
- lets release Thursday morning
PR Triage
other business
- Python 3? [adrian]
- webapp, yeah, maybe
- leeroy, well, probably hard
- configman, far away still
- what is our bus factor for AWS? [adrian]
- low risk because lonnen and rhelmer and phrawtzy are still "around"
- extract out socorro-collector and socorro-processors
- Terraform work, earlier this year, planned for this
- common-socorro library and connecting distinct repos by PyPI alike
- breaking out the collector would be "25 lines of code" --lars
- lars is not convinced :)
Travel, etc
- lars - out 11-4 through 11-13 for gettin' hitched.
- jp - 31st oct - 5th Nov Conference, 5th - 11th vacation