Closed Bug 833472 Opened 13 years ago Closed 11 years ago

Decommission or move to newer hardware vertica{1,2,3}.stage.metrics.scl3

Categories

(Infrastructure & Operations :: MOC: Service Requests, task)

task
Not set
minor

Tracking

(Not tracked)

RESOLVED WONTFIX

People

(Reporter: dumitru, Unassigned)

References

Details

(Keywords: spring-cleaning, Whiteboard: :Generic)

vertica{1,2,3}.stage.metrics.scl3 are currently marked as being in production. They run on old hardware (HP Generation 1 servers), and we have the following options: a) decommission them. b) move services to newer hardware. Please let us know what the decision is, so we can get the ball rolling. We recently renewed their warranty for one year only, so there's plenty of time to properly plan and schedule either options. Thanks!
Summary: Decommission or move to newer hardware for vertica{1,2,3}.stage.metrics.scl3 → Decommission or move to newer hardware vertica{1,2,3}.stage.metrics.scl3
We were planning on merging these machines into a cluster with the newer nodes (vertica{4,5,6}. If we wish to decommission the hardware, we should replace it with newer hardware equivalent to the newer nodes. That would likely involve ordering replacement hardware given the memory and disk layout.
(In reply to Daniel Einspanjer :dre [:deinspanjer] from comment #1) > We were planning on merging these machines into a cluster with the newer > nodes (vertica{4,5,6}. If we wish to decommission the hardware, we should > replace it with newer hardware equivalent to the newer nodes. That would > likely involve ordering replacement hardware given the memory and disk > layout. Yes, we'd like to decommission the hardware.
So we agreed to decommission and replace the hardware, who is responsible for ordering the replacements? We don't need to change the specs, just keep them roughly equivalent to vertica[4-6].
(In reply to Daniel Einspanjer :dre [:deinspanjer] from comment #3) > So we agreed to decommission and replace the hardware, who is responsible > for ordering the replacements? We don't need to change the specs, just keep > them roughly equivalent to vertica[4-6]. I can order the hardware. Would the exact same be sufficient? (newer hardware of course) or do you like to bump up RAM/HDD space etc?
no special upgrades needed, just roughly the same as vertica6
Flags: needinfo?(shyam)
Punting over to Dumitru. Please see if we can re-purpose something before we move to newer hardware. Also, probably hold on to this for a while before we take a call.
Assignee: server-ops → dgherman
Flags: needinfo?(shyam)
Would like to pick this up again. We need to get machines equivalent to vertica[4-6].metrics.scl3.mozilla.com to merge in to the cluster. Do we have anything suitable that we can re-purpose or do we need to order?
Assignee: dgherman → server-ops
Shyam what's the best way to find equivalent hardware so we can move on with this bug ?
Flags: needinfo?(shyam)
(In reply to Ludovic Hirlimann [:Usul] from comment #8) > Shyam what's the best way to find equivalent hardware so we can move on with > this bug ? Annie, do we still need a stage instance for vertica? If so, can these be virtualized or does it need to be hardware?
Flags: needinfo?(shyam) → needinfo?(aelliott)
This discussion needs to hold for the Vertica contract renewal discussions.
Flags: needinfo?(aelliott)
Whiteboard: :Generic
We're keeping Vertica, but we have vertica 4-6 in stage. Let's decom these. You can coordinate with mpressman, but I believe that nobody uses these.
Assignee: server-ops → server-ops-dcops
colo-trip: --- → scl3
mpressman: anything standing in the way of straight-up powering 1-3 off?
Flags: needinfo?(mpressman)
See Also: → 1064892
Greg - we're likely going to migrate away from Vertica in 2015, so let's keep stage alive until we're sure we're going to do that, as we may need to test some stuff out.
Flags: needinfo?(mpressman)
Group: infra
Component: Server Operations → MOC: Service Requests
Product: mozilla.org → Infrastructure & Operations
QA Contact: shyam → lypulong
I'm going to resolve/wontfix - we are OK with taking the risk of having old hardware on stage.
Status: NEW → RESOLVED
Closed: 11 years ago
Resolution: --- → WONTFIX
You need to log in before you can comment on or make changes to this bug.