This was going to be a test of hhvm, but we'll go straight to php7. Need to reinstall snapshot1001 to stretch, get php7 in the stretch manifests, make sure there's a buldof luasandbox and wikidiff, and give it a spin.
Description
Details
Status | Subtype | Assigned | Task | ||
---|---|---|---|---|---|
Resolved | Jdforrester-WMF | T172165 Require either PHP 7.0+ or HHVM in MW 1.31 | |||
Resolved | ArielGlenn | T181029 Upgrade dump hosts to stretch with php7 | |||
Resolved | ArielGlenn | T161509 Test php7 on snapshot1001 for dumps | |||
Resolved | ArielGlenn | T193399 Test all misc dumps cron jobs on snapshot1001 with php7/stretch |
Event Timeline
Script wmf-auto-reimage was launched by ariel on neodymium.eqiad.wmnet for hosts:
snapshot1001.eqiad.wmnet
The log can be found in /var/log/wmf-auto-reimage/201804240836_ariel_18873_snapshot1001_eqiad_wmnet.log.
Change 428577 had a related patch set uploaded (by ArielGlenn; owner: ArielGlenn):
[operations/puppet@production] use default installer for snapshot hosts
Change 428577 merged by ArielGlenn:
[operations/puppet@production] use default installer for snapshot hosts
Completed auto-reimage of hosts:
['snapshot1001.eqiad.wmnet']
and were ALL successful.
Change 428599 had a related patch set uploaded (by ArielGlenn; owner: ArielGlenn):
[operations/puppet@production] use php7.0 on snapshot1001 for dumps testing
Change 428599 merged by ArielGlenn:
[operations/puppet@production] use php7.0 on snapshot1001 for dumps testing
I've been running isolated jobs on the host and so far it's been doing the right thing. More testing tomorrow.
Running a couple of "pass over all wikis and do this job" scripts over there, should take a day for those to complete.
Change 429802 had a related patch set uploaded (by ArielGlenn; owner: ArielGlenn):
[operations/puppet@production] provide all scripts for dumps misc crons on testbed host
Change 429802 merged by ArielGlenn:
[operations/puppet@production] provide all scripts for dumps misc crons on testbed host
We see SET NAMES utf8mb4 in headers for mysql table dumps, rather than SET NAMES utf8. That's a side effect of going to stretch. I guess there could be some discussion in the future of what we want (binary, utf8mb4, something else), in any case, I don't see this as a stopper.