On Tue, 30 Apr 2013 08:37:44 -0500
Serge Hallyn <serge.hal...@ubuntu.com> wrote:
> Quoting Dwight Engen (dwight.en...@oracle.com):
> > +/*
> > + * Build the path to the log file
> > + * @name : the name of the container
> > + * @lxcpath : the lxcpath to use as a basename or NULL to use
> > LOGPATH
> > + * Returns malloced path on sucess, or NULL on failure
> > + */
> > +static char *build_log_path(const char *name, const char *lxcpath)
> > {
> > char *p;
> > int len, ret;
> >
> > /*
> > - * '$logpath' + '/' + '$name' + '.log' + '\0'
> > - * or
> > + * If USE_CONFIGPATH_LOGS is true the resulting path will
> > be:
> > * '$logpath' + '/' + '$name' + '/' + '$name' + '.log' +
> > '\0'
> > - * sizeof(LOGPATH) includes its \0
> > + *
> > + * If USE_CONFIGPATH_LOGS is false the resulting path will
> > be:
> > + * '$logpath' + '/' + '$name' + '.log' + '\0'
> > */
> > - len = sizeof(LOGPATH) + strlen(name) + 6;
> > + len = strlen(name) + 6; /* 6 == '/' + '.log' + '\0' */
> > + if (!lxcpath)
> > + lxcpath = LOGPATH;
> > #if USE_CONFIGPATH_LOGS
> > - len += strlen(name) + 1; /* add "/$container_name/" */
> > + len += strlen(lxcpath) + 1 + strlen(name) + 1; /* add
> > "/$container_name/" */ +#else
> > + len += strlen(lxcpath) + 1;
> > #endif
> > p = malloc(len);
> > if (!p)
> > return p;
> > #if USE_CONFIGPATH_LOGS
> > - ret = snprintf(p, len, "%s/%s/%s.log", LOGPATH, name,
> > name);
> > + ret = snprintf(p, len, "%s/%s/%s.log", lxcpath, name,
> > name); #else
> > - ret = snprintf(p, len, "%s/%s.log", LOGPATH, name);
> > + ret = snprintf(p, len, "%s/%s.log", lxcpath, name);
>
> Hi Dwight,
>
> I've pushed this patch to staging, but I do notice an oddity when
> USE_CONFIGPATH_LOGS=n (as on ubuntu). When I now start a container
> with an lxcpath, i.e. lxc-start -n c1 -P /home/serge/lxcbase,
> then the logfile is at /home/serge/lxcbase/dir1.log, not
> /home/serge/lxcbase/dir1/dir1.log. I think it would be better to have
> something like
>
> #if USE_CONFIGPATH_LOGS
> ret = snprintf(p, len, "%s/%s/%s.log", lxcpath, name, name);
> #else
> if (lxcpath == default_lxc_path())
> ret = snprintf(p, len, "%s/%s.log", lxcpath, name);
> else
> ret = snprintf(p, len, "%s/%s/%s.log", lxcpath, name,
> name); #endif
>
> Well, that or we use $lxcpath/logs/$name.log, but I prefer to keep the
> rule: we either use $LOGPATH/$name.log, or else
> $lxcpath/$name/$name.log, always.
>
> Thoughts?
Hi Serge, Hmm, I was just keeping the old behavior of
USE_CONFIGPATH_LOGS which meant you would always get the extra dir in
the middle (even on $LOGPATH), and otherwise the log would be in
$LOGPATH (or $lxcpath with this change).
I see your point though that having logs right in $lxcpath isn't so
nice, and I like what you've outlined above although I guess it
does slightly change the meaning of USE_CONFIGPATH_LOGS. I'll
submit a patch to do what you've suggested, and we can see if anyone
else cares to chime in. Thanks.
> thanks,
> -serge
------------------------------------------------------------------------------
Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET
Get 100% visibility into your production application - at no cost.
Code-level diagnostics for performance bottlenecks with <2% overhead
Download for free and get started troubleshooting in minutes.
http://p.sf.net/sfu/appdyn_d2d_ap1
_______________________________________________
Lxc-devel mailing list
Lxc-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/lxc-devel
