Still no root cause for 1424. We established that a small 100ms sleep solves the hanging problem on the MBA. One etap, 220-compaction-daemon.. seemed to continue to cause a hang on the MINI, but that looks to be an interaction with the newest performance patch and that's been fixed. Jan will confirm on the MINI. Details are in JIRA
My suspicion is that R15B is bringing out issues with how we shutdown things in couchdb, the hangs seem to always occur at the end of a given test, which explains why allowing etap:end_test() to wait a bit appears to clear it up completely. On Mar 8, 2012, at 5:55 AM, Noah Slater wrote: > Wow, cool. Thanks Bob, Filipe, and everyone who tested! :D > > Remaining issues: > > * Bob Dionne will driving COUCHDB-1424 > * Benoît Chesneau will be driving COUCHDB-1426 > > Guys? > > On Thu, Mar 8, 2012 at 10:19 AM, Robert Newson <[email protected]> wrote: > >> Filipe's patch restores (and then exceeds) performance from before the >> regression, the explanation is lucid, and no contradictory results >> have been observed since, so I think that issue is closed now. >> >> B. >> >> On 8 March 2012 01:08, Noah Slater <[email protected]> wrote: >>> Hey guys, >>> >>> What's the status of these items? >>> >>> On Sun, Mar 4, 2012 at 1:29 AM, Noah Slater <[email protected]> >> wrote: >>> >>>> Hey chaps, >>>> >>>> How are we doing on: >>>> >>>> * Bob Dionne will driving COUCHDB-1424 >>>> * Benoît Chesneau will be driving COUCHDB-1426 >>>> * Robert Newson will be driving the performance work >>>> >>>> The R15B patch has been committed already. >>>> >>>> Thanks! >>>> >>>> N >>>> >>
