About 16 instances (presumably with busted puppet) still think virt1000 is their salt master. They should get pushed over to labcontrol1001 before virt1000 is shut down.
Description
Status | Subtype | Assigned | Task | ||
---|---|---|---|---|---|
Resolved | Andrew | T90625 virt1000 SPOF | |||
Resolved | • Cmjohnson | T102646 Rename virt1000 to labcontrol1002, move to same subnet as labcontrol1001 | |||
Resolved | Andrew | T102005 Kill off virt1000 | |||
Resolved | Andrew | T103010 Sort out remaining virt1000 salt minions |
Event Timeline
It looks like some were missed somehow. I see these instances
cephtest-1.cephtest.eqiad.wmflabs
cephtest-3.cephtest.eqiad.wmflabs
cephtest-puppetmaster.cephtest.eqiad.wmflabs
graphite-jessie.graphite.eqiad.wmflabs
language-dev.language.eqiad.wmflabs
mobile-hierator4.mobile.eqiad.wmflabs
moses1.fatg.eqiad.wmflabs
openstack-test-ceilometer.openstack.eqiad.wmflabs
safesandbox.safesandbox.eqiad.wmflabs
test-carbon-c-relay.graphite.eqiad.wmflabs
which have virt1000 in their salt configs.
and they all have a puppet master other than the standard labs-puppetmaster-eqiad.wikimedia.org
also broken puppet on them all afaict.
I should mention they also have i-000-style names in their salt configs too.
Reopening for now.
Still three left for whatever reason.
android-build.mobile.eqiad.wmflabs
mobile-hierator2.mobile.eqiad.wmflabs
test-carbon-c-relay.graphite.eqiad.wmflabs