Thread: gitmaster -> anon git repo propagation broken?
Alvaro's recent commits (00f15338b234e5fd et al) appear on gitmaster but not on the anonymous repo, which means they are not getting tested on the buildfarm. This is not good on the day of a release. regards, tom lane
Tom, * Tom Lane (tgl@sss.pgh.pa.us) wrote: > Alvaro's recent commits (00f15338b234e5fd et al) appear on gitmaster > but not on the anonymous repo, which means they are not getting tested on > the buildfarm. This is not good on the day of a release. There was a failure in the push mechanism but it appears to have cleared and the two look current to me now. I have a feeling we need to review the push script a bit to not update the 'last push' timestamp until after we have *successfully* pushed, otherwise it gets "stuck" until another commit happens. Not going to change that now, given the timing, but will put it on my list of things to look at after the release (will plan to chat about it with other pginfra folks in Estonia). Thanks! Stephen
Stephen Frost <sfrost@snowman.net> writes: > * Tom Lane (tgl@sss.pgh.pa.us) wrote: >> Alvaro's recent commits (00f15338b234e5fd et al) appear on gitmaster >> but not on the anonymous repo, which means they are not getting tested on >> the buildfarm. This is not good on the day of a release. > There was a failure in the push mechanism but it appears to have cleared > and the two look current to me now. Looks good from here, too, thanks. regards, tom lane
Mon, Oct 24, 2016 at 3:42 PM, Stephen Frost <sfrost@snowman.net> wrote:
Tom,
* Tom Lane (tgl@sss.pgh.pa.us) wrote:
> Alvaro's recent commits (00f15338b234e5fd et al) appear on gitmaster
> but not on the anonymous repo, which means they are not getting tested on
> the buildfarm. This is not good on the day of a release.
There was a failure in the push mechanism but it appears to have cleared
and the two look current to me now.
I have a feeling we need to review the push script a bit to not update
the 'last push' timestamp until after we have *successfully* pushed,
otherwise it gets "stuck" until another commit happens. Not going to
change that now, given the timing, but will put it on my list of things
to look at after the release (will plan to chat about it with other
pginfra folks in Estonia).
I think the reasoning behind that is we need to touch it before in case a new commit shows up while we push it. I guess what we need to do is touch a tempfile somewhere and the mv it over the actual file if things succeed. We can probably use the existing lockfile.
But yes, let's shoot holes in that idea in Tallinn.
* Tom Lane (tgl@sss.pgh.pa.us) wrote: > I wrote: > > Stephen Frost <sfrost@snowman.net> writes: > >> * Tom Lane (tgl@sss.pgh.pa.us) wrote: > > Alvaro's recent commits (00f15338b234e5fd et al) appear on gitmaster > > but not on the anonymous repo, which means they are not getting tested on > > the buildfarm. This is not good on the day of a release. > > >> There was a failure in the push mechanism but it appears to have cleared > >> and the two look current to me now. > > > Looks good from here, too, thanks. > > Well, it looked good at the time, but I see 2dde01ccbfb4c53c hasn't > propagated yet either. Ugh. I've kicked it. Thanks! Stephen
I wrote: > Stephen Frost <sfrost@snowman.net> writes: >> * Tom Lane (tgl@sss.pgh.pa.us) wrote: > Alvaro's recent commits (00f15338b234e5fd et al) appear on gitmaster > but not on the anonymous repo, which means they are not getting tested on > the buildfarm. This is not good on the day of a release. >> There was a failure in the push mechanism but it appears to have cleared >> and the two look current to me now. > Looks good from here, too, thanks. Well, it looked good at the time, but I see 2dde01ccbfb4c53c hasn't propagated yet either. regards, tom lane