after the bmo/4.2 upgrade, bzapi is returning a 302 error for some queries. https://api-dev.bugzilla.mozilla.org/latest/bug/722597 works https://api-dev.bugzilla.mozilla.org/latest/bug/722597/comment fails i suspect it's getting the old ip from somewhere, which will return the hardhat redirect.

> https://api-dev.bugzilla.mozilla.org/latest/bug/722597 works > https://api-dev.bugzilla.mozilla.org/latest/bug/722597/comment fails For whatever it's worth, https://api-dev.bugzilla.mozilla.org/latest/bug/722597?include_fields=comments successfully returns the comment objects.

Since we closed the trees for the upgrade because we wouldn't be able to star intermittent failures, and we aren't able to still, trees are closed.

I can tell that this URL is trying to reach an address in phx1 that none of us here have ever heard of (webheads81???). Nor am I able to easily find where in the code it is trying to reach this address. Nor is there service documentation for it in our internal wiki. Nor is this service running on our centrally managed OS and stack. I'd recommend that the tree not depend on this service if at all possible. :(

(In reply to Corey Shields [:cshields] from comment #4) > I can tell that this URL is trying to reach an address in phx1 that none of > us here have ever heard of (webheads81???). Take this part back, the reverse was happening.

Alright.. The reason why this worked in allizom but not in production is because allizom had a bugfix specifically for this that prod did not have. I've updated prod. This URL works now. I'm still not happy with the state of this app overall. If we are going to depend on it moving forward it should be moved into our standard stack and properly deployed (versus the many directories to match individual versions on the server today, with 'latest' being a real directory itself).

I filed a bug on not using it, bug 652454. bteam was going to implement it as a supported tier-1 service on bmo, bug 747201. yourteam was going to move it out of the community vlan, bug 743916.

(In reply to Phil Ringnalda (:philor) from comment #7) > yourteam was going to move it out of the community vlan, bug 743916. Cool, thanks for linking that.. 743916 should be a wontfix - I don't want to support one-off ubuntu VMs. The app should be moved into an environment that we -do- support. Now that developer services has hired more help we should be able to make that happen in the near future.

And I have hopes that this 4.2 upgrade frees up the bugzilla devs to replace bzapi with native api. So we'll wait and watch.

Could I note in passing that, as far as I can see, nothing this bug reports was a problem with BzAPI itself, which continues to run just fine? It looks like there was a change management failure if allizom had a necessary fix which didn't make it to prod. I have no objections to BzAPI being a managed service, or becoming a Bugzilla extension, or whatever - but in the mean time, can people stop being rude about it when it's done nothing wrong? It has feelings too, you know... :-) Gerv

(In reply to Gervase Markham [:gerv] from comment #10) > Could I note in passing that, as far as I can see, nothing this bug reports > was a problem with BzAPI itself, which continues to run just fine? the fix from bug 826275 was committed on the /allizom/ endpoint, but not /latest/. once that was fixed, all has been well :)

(In reply to Byron Jones ‹:glob› from comment #11) > (In reply to Gervase Markham [:gerv] from comment #10) > > Could I note in passing that, as far as I can see, nothing this bug reports > > was a problem with BzAPI itself, which continues to run just fine? > > the fix from bug 826275 was committed on the /allizom/ endpoint, but not > /latest/. > once that was fixed, all has been well :) And that was a BzAPI issue... (not having the latest code in latest? :p)

(In reply to Phil Ringnalda (:philor) from comment #7) > bteam was going to implement it as a supported tier-1 service on bmo, bug > 747201. FWIW, this is a Q2 goal for the B(A)Team. dkl

<blush> I thought you were talking about the Bugzilla instances. OK, this is my mistake. I'll get down off the high ground. <shuffles feet> I didn't put 2 and 2 together and realise that if it were a problem on allizom, then fixing it in the allizom instance would not be sufficient for ever. /latest is now the development tip (that's not guaranteed to be the case; the rule is just that it may be anywhere between the most recent stable release and the development tip), and all instances of BzAPI pointing at b.m.o. have been updated to know that b.m.o. is now version 4.2. Gerv