Hi Ed, On Fri, 2009-03-27 at 14:16 -0400, Ed Cashin wrote:
> If vbladed wasn't already done, this issue would be more interesting, > but since the functionality is already present, it's probably only > worth discussing if there are some specific significant shortcomings > of vbladed. The fact that it uses the shell isn't a shortcoming in > any way that matters, as far as I know. Ah, ok, I didn't know it was more or less finalized. The point of my modifications started when I needed more noise to syslog, and using a different facility code (though the script could be modified to use whatever by passing an option to logger). As I thought about it, it just seemed easier to get rid of the script. I actually did add the daemon code, properized the help display and others. posix_fadvise() and mlockall() did help a little on older / slower nodes with 100M nics, but not significantly. But, that's the beauty of free software, we can adapt it to fit our specific needs. So, how about vbladed checking for vbladedrc and reading it if it exists? This lets someone easily modify the logger options, perhaps set a nice value, etc. For instance, LOG_LOCAL7 for me would go to a remote syslog server, so its easier to monitor events from all storage nodes in one place. If someone hacks the script, then updates their software .. well, it gets confusing. Cheers! --Tim ------------------------------------------------------------------------------ _______________________________________________ Aoetools-discuss mailing list Aoetools-discuss@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/aoetools-discuss
