Comments on: We have how many machines? (“dedicated specialised slaves” vs “pool of identical slaves”)

By: John O’Duinn’s Soapbox » Unittest and l10n moved from “dedicated specialized slaves” to “pool of identical slaves”

Mon, 16 Mar 2009 19:48:24 +0000

[...] That reduced the risk that we’d have to close the tree for machine failure of one build machine. It also made setting up builds on new branches like Tracemonkey relatively quick and relatively easy. (more details here). [...]

By: John O’Duinn’s Soapbox »

John O’Duinn’s Soapbox » — Fri, 10 Oct 2008 22:22:54 +0000

[...] If you recall from this earlier post, we’ve moved from having one dedicated-machine-per-build-purpose, to now using a pool of shared identical machines. Pending jobs/builds are queued up, and then allocated to the next available idle slave, without caring if its a opt-build, debug-build, etc. Any slave can do the work. More importantly, failure of any one slave does not close the tree. [...]

By: Robert Helmer

Robert Helmer — Fri, 16 May 2008 06:43:08 +0000

@Dave: “buildslave” is the Buildbot nomenclature for what John is talking about. “Builder” defines a role and “buildslave” is the machine chosen to fulfill a task. See http://buildbot.net/repos/release/docs/buildbot.html#System-Architecture

By: Dave

Dave — Fri, 16 May 2008 02:32:11 +0000

Distributed is a great way to go, just make sure you audit who did what so you don’t end up with a broken build you can’t trace back. Also I would suggest naming them something other than “slaves” perhaps “drones” or “bots” or just “builders”.

Comments on: We have *how* many machines? (“dedicated specialised slaves” vs “pool of identical slaves”)

By: John O’Duinn’s Soapbox » Unittest and l10n moved from “dedicated specialized slaves” to “pool of identical slaves”

By: John O’Duinn’s Soapbox »

By: Robert Helmer

By: Dave

Comments on: We have how many machines? (“dedicated specialised slaves” vs “pool of identical slaves”)