Someone's asked me to remove the old release candidates for 4.4, which seems like a good idea - the reporter said it looked confusing. So I've deleted all files in /var/www/files/4.4/ except for Squeak4.4-12327.zip.
Just in case you're wondering where those files have gone.
I'm happy with the idea that subsequent 4.4 releases go in that directory (and stay there), but a periodic cleanup of alpha or RC images seems sensible?
frank
On Tue, 1 Jan 2013, Frank Shearar wrote:
Someone's asked me to remove the old release candidates for 4.4, which seems like a good idea - the reporter said it looked confusing. So I've deleted all files in /var/www/files/4.4/ except for Squeak4.4-12327.zip.
Just in case you're wondering where those files have gone.
I'm happy with the idea that subsequent 4.4 releases go in that directory (and stay there), but a periodic cleanup of alpha or RC images seems sensible?
Releases belong to the release directory, while alphas/RCs belong to the matching alpha directory. For 4.4 these are
http://ftp.squeak.org/4.4/ http://ftp.squeak.org/4.4alpha/
Levente
frank
On 1 January 2013 23:33, Levente Uzonyi leves@elte.hu wrote:
On Tue, 1 Jan 2013, Frank Shearar wrote:
Someone's asked me to remove the old release candidates for 4.4, which seems like a good idea - the reporter said it looked confusing. So I've deleted all files in /var/www/files/4.4/ except for Squeak4.4-12327.zip.
Just in case you're wondering where those files have gone.
I'm happy with the idea that subsequent 4.4 releases go in that directory (and stay there), but a periodic cleanup of alpha or RC images seems sensible?
Releases belong to the release directory, while alphas/RCs belong to the matching alpha directory. For 4.4 these are
I've known about 4.4alpha, but last I heard we wanted RCs to go to 4.4 (because that's where I've been putting them for the last month or so).
At any rate, I've recorded this at http://wiki.squeak.org/6189.
Thanks!
frank
Levente
frank
I have a watchdog timer in my InterpreterVM job so that the Squeak image will exit if the job is not complete within some time period. I was starting to get failures on the builds as job runtime exceeded 10 minutes, so I bumped my timer up from 10 to 15 minutes, and all is well again.
Nothing to worry about, just FYI as we will need to keep an eye on total load as we add more processing to the system.
Dave
On 3 January 2013 18:43, David T. Lewis lewis@mail.msen.com wrote:
I have a watchdog timer in my InterpreterVM job so that the Squeak image will exit if the job is not complete within some time period. I was starting to get failures on the builds as job runtime exceeded 10 minutes, so I bumped my timer up from 10 to 15 minutes, and all is well again.
Nothing to worry about, just FYI as we will need to keep an eye on total load as we add more processing to the system.
It looks like it's perfectly possible to run slaves in a headless fashion, so that they can run behind NATs and connect _to_ the master. The angband node is an example of this. So scaling CI horizontally is easy, by people who can run slaves as and when they're able.
frank
Dave
box-admins@lists.squeakfoundation.org