Okay. I won't have time until tomorrow, I think. I already tore down the instance and its storage.
On Thu, Mar 14, 2013 at 12:08 PM, Jan Lehnardt <[email protected]> wrote: > Can you two post `couch.log`s and stdout/stderr for the runs with failures? > > Cheers > Jan > -- > > > On Mar 14, 2013, at 00:31 , Randall Leeds <[email protected]> wrote: > >> I took some time today to run the 1.3.x branch through the tests on an >> EBS-optimized large instance running the 64-bit RHEL 6.3 AMI with a >> 10GB XFS partition mounted as /usr/local/var/lib/couchdb. >> >> Here's my experience and notes on what's different from the current >> RHEL5 installation wiki page: >> >> - libcurl from the repositories is new enough now >> >> - gcc-g++ is now a requirement, for snappy >> >> - I had to merge COUCHDB-1692 [0] and COUCHDB-1699 [1] in order to >> successfully build from git source >> >> Here's what happened with the test suite: >> >> - First time segfault. Logs about local.ini. After chown >> /usr/local/etc/couchdb it seems to work. >> - form_submit failed the first time. Succeeded when run individually >> - attachment_ranges fails in Chrome but not FF. I think this is expected >> IIRC. >> - stats failed the first time and said it managed to trigger an all >> dbs active, worked second time >> - replication and replicator_db did pass first try (after 6 minutes >> and 2 minutes respectively) >> >> I propose we merge the patches in the aforementioned tickets (and the >> init script change in COUCHDB-1449 [2]) to master and 1.3.x. >> >> All in all, it looks pretty good, actually. >> >> -Randall >> >> [0] https://issues.apache.org/jira/browse/COUCHDB-1692 >> [1] https://issues.apache.org/jira/browse/COUCHDB-1699 >> [2] https://issues.apache.org/jira/browse/COUCHDB-1449 >
