On Apr 18, 2013, at 10:25 AM, Bryan Call <bc...@yahoo-inc.com> wrote:
> Couldn't resolve the origin hostname. I had to turn on debugging in traffic > server and saw that it couldn't resolve the hostname for the origin server. > I don't remember if the response to the client was an empty reply or if there > was a valid http error response. Ick, that's nasty. We should make sure that it auto-upgrades hostDB for people coming from 3.2 to 3.4. I filed https://issues.apache.org/jira/browse/TS-1832 J > > -Bryan > > On Apr 16, 2013, at 4:09 PM, James Peach <jpe...@apache.org> wrote: > >> On Apr 16, 2013, at 3:12 PM, Bryan Call <bc...@yahoo-inc.com> wrote: >> >>> +1 >>> >>> Passed all signatures, compiled, and passed regression. It is running as >>> my forward proxy now. One problem I ran into was it didn't work with a >>> hostdb size of 32M and I had to increase it to 200M. >> >> What were the symptoms of hostdb not working? >> >>> >>> -Bryan >>> >>> On Apr 16, 2013, at 8:31 AM, Leif Hedstrom <zw...@apache.org> wrote: >>> >>>> On 4/16/13 4:19 AM, Igor Galić wrote: >>>>> >>>>> - >>>>> make[3]: Leaving directory >>>>> `/tmp/ts/trafficserver-3.3.2-dev/iocore/eventsystem' >>>>> make check-TESTS >>>>> make[3]: Entering directory >>>>> `/tmp/ts/trafficserver-3.3.2-dev/iocore/eventsystem' >>>>> ERROR: unable to access() TS_ROOT '/opt/ats-3.3': 2, No such file or >>>>> directory >>>>> >>>>> /bin/bash: line 5: 19607 Segmentation fault (core dumped) ${dir}$tst >>>>> FAIL: test_Buffer >>>>> ERROR: unable to access() TS_ROOT '/opt/ats-3.3': 2, No such file or >>>>> directory >>>>> >>>>> /bin/bash: line 5: 19631 Segmentation fault (core dumped) ${dir}$tst >>>>> FAIL: test_Event >>>>> ============================================= >>>>> 2 of 2 tests failed >>>>> Please report to dev@trafficserver.apache.org >>>>> ============================================= >>>> >>>> I have some fixes for this, which I will commit for v3.3.3. If everyone is >>>> ok with it, I'd prefer not to reroll v3.3.2, these fixes doesn't affect >>>> runtime or compile time at all (only make test fails, which is bad, but >>>> not >>>> mission critical). The changes does not change anything in the core code, >>>> only in the building of the test programs themselves. With my fixes, all >>>> tests succeeds, for v3.3.2 as well. >>>> >>>> Thanks Igor. >>>> >>>> -- Leif >>>> >>> >> >
