(sending from the right email account now, first attempt will probably
bounce ...)
Hi All,
currently response of QGIS website and Issue trackers are not too good
(understatement).
if I understand the discussion correctly there is a hardware problem
which is hard to solve?
is it maybe wise to move QGIS website(s) to our other server for the
time being?
For the issue-tracker that is not that easy, though there is discussion
about the possibility to Github-issues, so my idea is to at least the
websites away. The rsyncing of the output takes a lot of time (3 pdf
files: 4 minutes, whole doc site: 20 minutes), so maybe we can take the
burden from the server in that way.
Is it possible to have a raw time indication for the hardware fix, that
would make our decision to move easier.
just curious, are you also on the "osgeo4" hardware?
I thought you had a Hetzner server...
Best
Markus
Yes qgis.org and hub.qgis are on osgeo4. This machine had a disk fail on
it's Raid 6. Theoretically that shouldn't degrade the performance, but
maybe that's what happened.
We are ordering a replacement.
Also I spoke with OSUOSL and we are devising a plan this week to buy a
new server with the aim to fix the long standing issues with osgeo4.
just curious, are you also on the "osgeo4" hardware?
I thought you had a Hetzner server...
Hi Markus,
nope the web-stuff: site, docs, debian packages are still on osgeo
hardware, check: http://hub.qgis.org http://qgis.org
to see the slowness I mean
about one year ago we moved plugins.qgis.org and all the building
machinery (packages and sphinx for docs and site) to a Hetzner server
indeed.
That's why I'm asking for a planning, as currently the site is too slow
for normal use, and also the rsyncing from Hetzner ('qgis2') to osgeo
('qgis') takes too long.
If it takes a week to solve the hardware problem, we probably wait for
it, but it is more I think we should move the site temporarily to qgis2
(also because during a release: around 20th feb we generate a lot of
traffic).
if anybody can provide a planning that would be easier for us to decide
what to do.
On 02/03/2015 08:47 AM, Richard Duivenvoorde wrote:
On 03-02-15 11:20, Markus Neteler wrote:
Hi Richard,
just curious, are you also on the "osgeo4" hardware?
I thought you had a Hetzner server...
Hi Markus,
nope the web-stuff: site, docs, debian packages are still on osgeo
hardware, check: http://hub.qgis.org http://qgis.org
to see the slowness I mean
about one year ago we moved plugins.qgis.org and all the building
machinery (packages and sphinx for docs and site) to a Hetzner server
indeed.
That's why I'm asking for a planning, as currently the site is too slow
for normal use, and also the rsyncing from Hetzner ('qgis2') to osgeo
('qgis') takes too long.
If it takes a week to solve the hardware problem, we probably wait for
it, but it is more I think we should move the site temporarily to qgis2
(also because during a release: around 20th feb we generate a lot of
traffic).
if anybody can provide a planning that would be easier for us to decide
what to do.
Regards,
Richard Duivenvoorde
I would recommend, you flip the main qgis.org over to your backup on
qgis2 for now. Hopefully diverting that traffic will leave hub
performing better as is.