[tahoe-dev] [tahoe-lafs] #368: munin pages on .org have broken links to dev pages
tahoe-lafs
trac at tahoe-lafs.org
Fri Aug 6 23:08:47 UTC 2010
#368: munin pages on .org have broken links to dev pages
-------------------------+--------------------------------------------------
Reporter: terrell | Owner: somebody
Type: defect | Status: new
Priority: minor | Milestone: undecided
Component: website | Version: 0.9.0
Resolution: | Keywords: munin website
Launchpad Bug: |
-------------------------+--------------------------------------------------
Comment (by freestorm):
Is this ticket still alive?
from IRC:
{{{
<FreeStorm> warner: have you 2min for ticket #368?
<FreeStorm> just to be sure I understand what is the configuration and the
issue
* warner looks
<warner> hm, I'd completely forgotten about those graphs
<FreeStorm> you have 2 munin-node, 1 "private" , and 1 "public" ?
<FreeStorm> And you want to show only some graph from the "private" to the
"public" right?
<warner> I'm not really sure any more
<warner> we used to have a number of tests running on the buildbot, which
emitted numbers, which were then read by a munin plugin on the buildmaster
machine, and produced graphs on that machine
<warner> these graphs were then copied (via a cronjob and scp) from the
buildmaster machine (aka "hanford" aka "dev.allmydata.org") onto the
world-visible web server / trac machine (aka "org" aka "allmydata.org")
<warner> those tests depended upon many of the "testgrid" machines, so I
suspect that none of them are running anymore
<FreeStorm> Okay, buildbot=private allmydata.org=public
<warner> right
<FreeStorm> SO I will write a comment, maybe someone know more how it
supposed to work, and if this ticket need to be close
<warner> I suspect the ticket should be closed and a new one opened named
"buildbot-driven performance/coverage munin graphs should be revived"
<FreeStorm> with munin, you can display munin-graph from one machine to
another, but I think you cannot chose which graphy, maybe it's only all or
nothing
<warner> because the problem is now much deeper than merely fixing the
copy-graph-from-hanford-to-org problem
<warner> well, a given munin master will generate a bunch of .png files
under /var/www/munin/
<warner> then you can either serve them from that machine, or rig up a
cronjob like we did to copy them elsewhere
<warner> for allmydata, that "hanford" machine had a bunch of internal
graphs that we didn't want to publish (stuff about the production grid,
internal machines, etc)
<warner> so we were only copying a subset of the images, rather than
making hanford's webserver world-readable
<FreeStorm> yes, by this way you can chose what you want to display.
<warner> our general goal these days is to move things away from hanford
and onto org, so we can retire hanford eventually
<warner> that means moving the buildmaster to org, then rebuilding the
testgrid structure enough to bring back the performance tests, then
rebuilding those graphs
<FreeStorm> Okay, I will write your comments to the ticket.
<warner> thanks!
}}}
--
Ticket URL: <http://tahoe-lafs.org/trac/tahoe-lafs/ticket/368#comment:4>
tahoe-lafs <http://tahoe-lafs.org>
secure decentralized storage
More information about the tahoe-dev
mailing list