FWIW configuration niceties are enough to sway me. I can't imagine either of them would perform poorly enough to make a difference for us.
On Tue, Dec 29, 2015 at 4:35 PM, John Sirois <j...@conductant.com> wrote: > On Tue, Dec 29, 2015 at 5:18 PM, John Sirois <j...@conductant.com> wrote: > > > > > > > On Tue, Dec 29, 2015 at 5:05 PM, John Sirois <j...@conductant.com> > wrote: > > > >> > >> > >> On Tue, Dec 29, 2015 at 5:02 PM, Jeff Schroeder < > >> jeffschroe...@computer.org> wrote: > >> > >>> Primarily it is faster, uses less memory, and annotates tracebacks with > >>> package versions. The last one seems like a winner for debugging user > >>> issues or operationally. > >>> > >>> http://logback.qos.ch/reasonsToSwitch.html > >>> > >>> I'm not strongly opinionated either way, but it does seem like a better > >>> log4j. > >>> > >> > >> Looks like this decision is nicely limited to a build.gradle edit: > >> http://logback.qos.ch/reasonsToSwitch.html#slf4j > >> > > > > After a brief skim of the configuration docs [1], I'm in favor of > > switching in a follow-up RB to https://reviews.apache.org/r/41777/ > > In short - logback supports pointing to a non-root config file via a > > system-property out of the box, this makes aurora a non-nuisance for > > operators, they can easily modify init scripts to point to a custom > config. > > > > [1] http://logback.qos.ch/manual/configuration.html > > > > Ah yes, easier said than done since we have /logconfig [1][2]. > Jeff - do you feel strongly enough about this to file an issue to > investigate / prove out perf wins / send up a change? (doing any part of > this or all of this would wonderful and I'd be happy to review). > > [1] > > https://github.com/apache/aurora/blob/master/commons/src/main/java/org/apache/aurora/common/net/http/handlers/LogConfig.java > [2] > > https://github.com/apache/aurora/blob/master/src/main/java/org/apache/aurora/scheduler/http/JettyServerModule.java#L221 > > > > > > >> > >>> On Tuesday, December 29, 2015, Bill Farner <wfar...@apache.org> wrote: > >>> > >>> > I don't have a strong opinion about logback vs log4j. Can you > >>> summarize > >>> > some of the tradeoffs? > >>> > > >>> > On Tue, Dec 29, 2015 at 3:52 PM, Jeff Schroeder < > >>> > jeffschroe...@computer.org <javascript:;>> > >>> > wrote: > >>> > > >>> > > What about using logback instead of log4j? It has some interesting > >>> > benefits > >>> > > over log4j and we wouldn't be the first large mesos framework to > >>> switch > >>> > to > >>> > > it. > >>> > > > >>> > > Personally, I'd love to see glog burn and die in a fire. > >>> > > > >>> > > On Monday, December 28, 2015, Bill Farner <wfar...@apache.org > >>> > <javascript:;>> wrote: > >>> > > > >>> > > > We're currently using some logging scaffolding carried over from > >>> > Twitter > >>> > > > commons. I would like to propose that we dismantle some of this > in > >>> > favor > >>> > > > of more standard java application logging conventions. > >>> > > > > >>> > > > Concretely, i propose we remove the following scheduler command > >>> line > >>> > > > arguments: > >>> > > > -logtostderr > >>> > > > -alsologtostderr > >>> > > > -vlog > >>> > > > -vmodule > >>> > > > -use_glog_formatter > >>> > > > > >>> > > > Instead of these, we can allow users to customize logging via > >>> standard > >>> > > > java.util.logging inputs (e.g. logging.properties). We could > >>> explore > >>> > > using > >>> > > > an alternative to java.util.logging, but i suggest we retain that > >>> > backend > >>> > > > for now (since it's what we're currently using). > >>> > > > > >>> > > > >>> > > > >>> > > -- > >>> > > Text by Jeff, typos by iPhone > >>> > > > >>> > > >>> > >>> > >>> -- > >>> Text by Jeff, typos by iPhone > >>> > >> > >> > >> > >> -- > >> John Sirois > >> 303-512-3301 > >> > > > > > > > > -- > > John Sirois > > 303-512-3301 > > > > > > -- > John Sirois > 303-512-3301 >
