scans keep timing out
by Max Spevack
I keep getting a simple timeout error on the scans that I run. I've
done it inside the VPN, outside the VPN, etc. Everything just works
fine, and then somewhere about 1/3 of the way through the lists, it
errors and bails out on the whole program.
The problem here is not a simple host unreachable error, but that the
recovery is absent, which means that it doesn't try to keep going and
connect again, and when I restart, it doesn't realize "hey, I just
scanned 25 out of 75 lists (which took me 2 hours) and I should start
where I left off." It starts again from the beginning.
I wonder if there is anything we can do to fix this, since it's actually
preventing us from collecting new data.
/usr/lib/ruby/1.8/net/http.rb:560:in `initialize': No route to host -
connect(2) (Errno::EHOSTUNREACH)
from /usr/lib/ruby/1.8/net/http.rb:560:in `open'
from /usr/lib/ruby/1.8/net/http.rb:560:in `connect'
from /usr/lib/ruby/1.8/timeout.rb:53:in `timeout'
from /usr/lib/ruby/1.8/timeout.rb:93:in `timeout'
from /usr/lib/ruby/1.8/net/http.rb:560:in `connect'
from /usr/lib/ruby/1.8/net/http.rb:553:in `do_start'
from /usr/lib/ruby/1.8/net/http.rb:542:in `start'
from /usr/lib/ruby/1.8/open-uri.rb:242:in `open_http'
... 20 levels...
from ./scanner.rb:66:in `each'
from ./scanner.rb:66:in `scan_lists'
from ./scanner.rb:42:in `run'
from ekg.rb:71
15 years, 5 months
Updated .gitignore
by Steve Milner
Added *.db and v2/graphs.html to the list. Yeah, I'm that lazy :-).
Steve 'Ashcrow' Milner
15 years, 5 months
prefix/postfix patches (light templates)
by Steve Milner
0001-move-html-files-into-a-templates-setup.patch
0002-allow-for-configurable-prefix-postfix-templates.patch
First patch moves around the template files ... second one actually
makes templates configurable and adds a second set that just uses a slightly
different color scheme :-P.
Steve 'Ashcrow' Milner
15 years, 5 months
Graph suggestions and display ideas.
by Michael DeHaan
I think right now EKG data is a bit misleading as it's hard to tell
"there are a lot of unique posters from this domain", versus 'there are
a lot of vocal posters using this domain".
Unique posters by domain over time, at least for the major domains
(fedoraproject, gmail, redhat.com, etc) seems like a good graph to
display as an option to default graph.
We could also play some nice CSS games to show one or the other but not
both at the same time and also be able to select which graphs to show in
the list.
I have been slack at EKG recently though I may chase this tomorrow.
If we're putting EKG on cron i'm not sure postgres is a huge concern but
that is still on my list to benchmark.
--Michael
15 years, 5 months
Four Quick Patches
by Steve Milner
0001-Trap-GetoptLong-InvalidOption-and-exit-with-it-s-ow.patch
0002-added-license-information-to-all-first-party-ruby-fi.patch
0003-removed-extra-whitespaces-at-the-end-of-files.patch
0004-added-.gitignore.patch
All four attached to this email.
Cheers,
Steve 'Ashcrow' Milner
15 years, 5 months
Google Chart and Fedora OSS policy
by Michael DeHaan
Currently I like the Google charts graphing pretty well -- it looks good
and produces no load on the client -- unlike my previous javascript
generator.
If we intend to ever host EKG on Fedora infrastructure we would need to
supply an alternate graphing system. This is not something I'm
presently looking for.
Though if we do want to find a sever to run EKG stats on, we need to
resolve this.
I am inclined to just keep working on the code in general and running
stats on our machine for a few weeks as replacing the grapher is not
super-happy-fun to me, though if someone else wants to handle this, that
would be fine too.
Thoughts?
--Michael
15 years, 5 months
EKG Friday -- postgres benchmarking
by Michael DeHaan
Friday is my "work on EKG some" day... I'm planning on trying to
benchmark some other databases and speedups unless anyone else has
something pressing.
Sound good?
--Michael
15 years, 5 months