[smokeping-users] Something Odd With This These Graphs: http://oss.oetiker.ch/smokeping-demo/?target=Customers.OP.johan
bhoule at siliconexus.com
Wed Oct 5 17:36:38 CEST 2011
SmokePing installation is a bear. But while I completely agree the RPM
approach certainly would have helped my own painful install, I just
don't get the tactic of saying "it was easier to do 2.4.2" and then
picking apart the old code.
This is an open-source tool for _any_ platform, not just for LSB. The
caretakers probably have enough to do with source, they can't also
maintain installable packages for all platform types. That takes the
community. Barring that, any streamlining from the developer side such
as Makemaker, autoconf, or Configure is much appreciated.
Reliance on RPM alone has pitfalls. I was surprised to find that the
"current" Apache RPM is by no means current. And this is for a
ubiquitous tool! RPM may be "easy", but its not always the holy grail
On 10/5/2011 3:05 AM, Jenny Lee wrote:
> > > Now the problems:
> > >
> > > a. What is the fix for this issue? I don't believe everyone is
> > > using it like that. The purpose of graphing... is so that we do
> > > not need to read the fine print. When I looked at that graph, I
> > > was to have a heart attack. How can I show something like this to
> > > my boss? I use cacti as well and I don't have this problem.
> > > Nothing is clear from these graphs beyond a day.
> > you should, because loss is BAD, it should not happen, hence the
> > choice of consolidation function ...
> So there is not a single functional host out of the 30-40 listed on
> the demo site? Because all of them are FULL BLUE in 400 days graphs.
> Just because there is 1% loss in 1 out of the 144 probes run does not
> mean that entire day should be marked as having packet loss.
> There is some flaw here and those graphs surely would look bad to
> anyone who looks at them. But just 400 packets out of the 6 Million or
> so were dropped in fact.
> Is there any way I can fix those by taking aggregates or averages or
> something else?
> > > b. The UI needs serious improvements. 70% of my screen is empty.
> > > Why not to plug in 4 graphs there?
> > I am planning on moving smokeping to extopus, see www.extopus.org
> > if you are in a position to provide resources for this project, let
> > me know ...
> It has potential. Unfortunately, I am just a worker.
> > > c. What is Speedy CGI? I had hard time trying to find it let
> > > alone being able to install it?
> > smokeping 2.6.x does not need speedycgi anymore ... it uses
> > fastcgi ... where did you see the speedycgi requirement ?
> SpeedyCGI requirement was still there in 2011 even though last version
> of that software was abandoned 9 years ago! It is still coming with
> the Fedora SmokePing RPM released in 2011.
> > > e. I had to install 42 rpms to get this working! I had to take 3
> > > srpms from Fedora and recompile them on Redhat 6. Take Fedora
> > > SmokePing RPM, recompile on Redhat. Yet, I still have version
> > > 2.4.2. I spent 2 days. This doesn't have to be so painful!
> > you should not be using an old version ... bad for the karma ...
> > 2.6.x has autoconf support ...
> Tobi, very few people run software with autoconf any more. I did run
> it. And it installed about 200 dependencies (thankfully to
> /opt/smokebin). I just needed Config::Grammar!
> I had to edit one file. I had to edit another. I had to modify some
> more. I had to change permissions... Then I had to write an initscript
> to start it... That is why we have RPM to take care of these stuff for
> us. It quickly got tiring so I powered off the non-persistent VM and
> installed 2.4.2 from Fedora RPMs.
> RPM approach is something that should be considered for widespread
> adoptation of sofware.
> smokeping-users mailing list
> smokeping-users at lists.oetiker.ch
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the smokeping-users