What kind of development are you worried about stalling?
Ideally, bug fixes shouldn't involve such a severe change that we have to redo the long running tests. I would think that if such a fix did happen it would be for a problem severe enough to warrant delaying the release. On Fri, Apr 4, 2014 at 10:41 AM, Josh Elser <[EMAIL PROTECTED]> wrote:
It's ultimately up to the Release Manager if something should go into 1.6.0, but I don't see why we wouldn't include bug fixes that don't invalidate tests.
IMHO, we have way too many development branches as is so in the case you're describing I'd rather have things wait in 1.5 instead of adding another branch. On Fri, Apr 4, 2014 at 11:31 AM, Josh Elser <[EMAIL PROTECTED]> wrote:
Can we circle back on where we are wrt required testing on 1.6.0? As is, there's < 24hr remaining on the RC1 vote.
I think we have outstanding:
1) Continuous Ingest run 24 hrs w/o agitation, verified
2) RandomWalk with LongClean for 24hrs
3) RandomWalk with LongClean and agitation for 24hrs
I believe Mike D. finished the 72hr CI w/o agitation and Bill H. has a #2 in progress.
I had #1 queued for this previous weekend on BigTop 0.7.0, but my cluster had NIC issues under stress and blew up. It did get to do some ingestion and it did make attempts for hte full 24hr. so if it gets fixed in time I still could go through recovery and verification.
Does anyone else have tests in progress that will fulfill #1 and #3? On Thu, Apr 10, 2014 at 1:40 PM, Bill Havanki <[EMAIL PROTECTED]>wrote: Sean
#2 is still going as of this morning, so I'm optimistic with less than 2 hours remaining that it will survive the 24 hours. However, I have nothing going for #3.
Josh has -1'd rc1. If there is no #3 running I'd -1 it, only because we ran out of testing time.
If no one else is running #3, I'd be happy to, after updating my cluster to the very latest (I'm about 5 days behind). Assuming rc1 doesn't ship, before testing I'd also like to commit ACCUMULO-2621 (about concurrent rw) and somehow deal with ACCUMULO-2657 (about shard rw). On Mon, Apr 14, 2014 at 7:10 PM, Sean Busbey <[EMAIL PROTECTED]> wrote: // Bill Havanki // Solutions Architect, Cloudera Govt Solutions // 443.686.9283
I'm running RW with ShortClean, and I have reported a number of issues (ACCUMULO-2655). They may be issues with my test environment, or the mechanism used to tear down the tests after they timeout. I'm still investigating.
On Tue, Apr 15, 2014 at 9:06 AM, Keith Turner <[EMAIL PROTECTED]> wrote:
My run of LongClean randomwalk just passed the 24-hour mark.
commit 63ca9c9 with patches for ACCUMULO-2621 7-node cluster on CDH 4.5 (Hadoop 2.0.0+CDH4.5.0), 2 masters, 5 tservers, 3 ZKs CentOS 6.4 64-bit
I'll start to update and re-tool for an agitation run. On Tue, Apr 15, 2014 at 9:41 AM, Eric Newton <[EMAIL PROTECTED]> wrote: // Bill Havanki // Solutions Architect, Cloudera Govt Solutions // 443.686.9283
I just started LongClean rw with agitation. (I only use one walker, unlike Eric who apparently uses many.) On Tue, Apr 15, 2014 at 11:18 AM, Eric Newton <[EMAIL PROTECTED]> wrote: // Bill Havanki // Solutions Architect, Cloudera Govt Solutions // 443.686.9283