Latest update on the issue: I ran the tests in a highly rigged environment with a classpath full of RC0 Hadoop 0.22 artifacts (instead of the snapshot ones from Maven). All of the tests pass now. I think this gets HBase/Hadoop RCs off the hook.
Bigtop is now +1 on HBase 0.92RC/Hadoop 0.22RC Thanks, Roman. P.S. On a related note -- I really starting to *hate* how unreliable snapshots from Apache projects are. This is really bad and makes me think that perhaps, we'd be better off having a Bigtop-specific maven repo to which I can push artifacts when doing stack builds. So at least, Bigtop tests will be unaffected by something like this (sorry folks, I don't know how to help you in general case of mvn test, etc.) P.P.S. On the other hand, that's yet another piece of infra. we'd have to maintain :-( On Mon, Dec 5, 2011 at 3:45 PM, Jonathan Hsieh <[email protected]> wrote: > Hey Stack, > > I won't be able to take a look at these until Wednesday. > > Jon. > > On Mon, Dec 5, 2011 at 3:42 PM, Stack <[email protected]> wrote: > >> On Mon, Dec 5, 2011 at 3:32 PM, Jonathan Hsieh <[email protected]> wrote: >> > Roman and I did some commit auditing on 0.92 and found that there are >> > several patches in the 0.90 branch but not in the 0.92 branch or release >> > candidate. I haven't looked at any of these deeply yet but I'd think we >> > should deal with these regressions before 0.92 goes out. >> > >> >> Let us know if you are going to look into it soon Jon, else I'll take a >> gander. >> St.Ack >> >> (Thanks for doing this) >> > > > > -- > // Jonathan Hsieh (shay) > // Software Engineer, Cloudera > // [email protected]
