On 27 April 2017 at 11:37, Sam Ruby <[email protected]> wrote: > On Thu, Apr 27, 2017 at 5:23 AM, sebb <[email protected]> wrote: >> site-check.rb does not check the site, and the output is redirected to >> site-scan.json. >> >> IMO it would make more sense to call it site-scan.rb. >> >> Agreed? >> >> Unfortunately this means changing the cron job, but we could get round >> the timing issues by creating temporary copy/alias until puppet has >> run. Or we could change puppet and wait for it to run, then change the >> script. This would result in some temporary failures. > > WFM. > > Temporary failures don't concern me - this isn't mission critical. > > Also note: it is the puppet run that would both change the cron job > and deploy the code.
Yes and no. The code and puppet cron job are in separate Git repos, and normally the puppet repo needs two people to make updates, which makes it harder to synch changes. However in this case it was possible to change the puppet repo directly. > And given that the cron job is hourly, it will > typically run thirty minutes later than the deployment. > > I've pushed these changes: > > https://github.com/apache/whimsy/commit/b3e4b3fd08d79b34a1dc80e7c8241aaf62952b41 > https://github.com/apache/infrastructure-puppet/commit/bcb4fab46496b5bf9f4b5e57b11ef1821ef045f0 Thanks! > - Sam Ruby
