Network measurement has been my bag - particularly in data center networks; RTTs are almost always submillisecond.
All the best everyone! - Rob . On Wed, Aug 5, 2020 at 6:31 PM Judah Milgram <milg...@cgpp.com> wrote: > > Thanks, makes sense. I just point ntpd to pool.ntp.org and take whatever > I get, and so far so good. And typically shut down up to a day at a time > and no ill effects. But discrete time adjustments do sound problematic. > Who actually needs sub-millisecond accuracy - database servers? Network > filesystems? > > > On 8/5/20 7:53 PM, Rob Sherwood wrote: > > If you don't have a permanent internet connection or access to a quality > > time server, there's no reason to run ntpd. Also, depends on what level > > of time precision your application(s) need, but ntpd causes little skips > > in time which can cause problems if you need sub-millisecond accuracy. > > > > HTH, > > > > - Rob > > . > > > > On Wed, Aug 5, 2020 at 4:21 PM Judah Milgram <milg...@cgpp.com > > <mailto:milg...@cgpp.com>> wrote: > > > > Ignorant question: > > > > Slackware ships with ntpd, but /etc/rc.d/rc.ntpd installs without > > execute permission, so one must deliberately decide to enable it. > > > > My question is: why would anyone *not* want to run ntpd at least as a > > client? OK, some will prefer chrony, but in general, is there any > > reason > > not to run a time server daemon? Let's assume it either accepts no > > incoming connections or is properly configured to feed time only to > the > > local network. > > > > Background: I've always done the standalone hwclock --adjust thing > and > > it never really worked well. So I tried ntpd. Big improvement. > > > > Hope everyone's staying safe and not too bored! > > > > Judah > > > > > > > > -- > > Judah Milgram > > milg...@cgpp.com <mailto:milg...@cgpp.com> > > 301-257-7069 > > > > You received this email because you are subscribed to the UM Linux > > User's Group (UM-LINUX) mailing list. If you would like to > > unsubscribe from this list, simply send an email to > > lists...@listserv.umd.edu <mailto:lists...@listserv.umd.edu> with > > the message signoff UM-LINUX in the body. > > > > > -- > Judah Milgram > milg...@cgpp.com > 301-257-7069 > You received this email because you are subscribed to the UM Linux User's Group (UM-LINUX) mailing list. If you would like to unsubscribe from this list, simply send an email to lists...@listserv.umd.edu with the message signoff UM-LINUX in the body.
