Title: Message Title
Kenneth Barber updated an issue
PuppetDB / PDB-1079
investigate moving acceptance to vmpooler
Change By:
Kenneth Barber
Story Points:
13
Investigatethevalueofmovingouracceptanceteststovmpooler.ThisisactuallyacloneofPDB-7,whichweclosedearlierduetotheinstabilityofvsphereatthetime.InthepastwehavebeenreluctantduetocapacityissueswithvsphereversusEC2(whichhasnone),we'dneedtodeterminethepotentialcapacitywewouldrequirebeforeevenconsideringthisIpresumeandmakesurewedon'tjustintroduceaproblemtotheVMclusterbybringingoverourtests.Forexampleinthepasttheplatformteamhasbeenaskedtoreducecelltesting,wewouldn'twanttoputourselvesinthissituationifwecanavoidit,asitputsusinaworseplacethantoday.DiscussionswithOps/Qengwouldberequiredtoestimatethis,andensuretheproperhardwareisinplacebeforecommencing.Sincewehavesomeverylargecelltests,thisisimportant.Alsostabilityhasbeenaconcernwiththatsolutionovertheyears,butIbelievetosomeextentthishasbeenimprovedovertime.Thatis,Idon'tthinkitsasmuchofanissueasitwas,butweshoulddiscusswiththeotherlargerteams(likethepuppetcoreteams)tounderstandthis.Theimageswoulddifferslightlyalso(especiallyolderimages),soI'mguessingchangestotheacceptancetestsetuproutineswouldneedtobemadepotentially,buttestingwouldsurfacethis.Thepipelineswouldneedchangingtocaterforthischange,plustheacceptanceteststodayhavealotofS3repositoryassumptions,asopposedtoobtainingpackagesfromthebuildsystemdirectly,sothatwouldneedtochange.InparticularforthepipelinetheS3stuffwouldneedtogo.ThisismoreeffortthanitseemsonthesurfaceIbelieve,soI'mputtinginahigh#ofstorypointstomakeitclearthisispotentiallynon-trivial.
Add Comment
This message was sent by Atlassian JIRA (v6.3.10#6340-sha1:7ea293a)