Since Martin wanted more details, here's the grand scheme for the next
few months. I'll probably try to call a meeting in October to work out
the details.
1. Put up a new download server at OSUOSL
2. Migrate trac to 0.12
3. Migrate Drupal to OSUOSL Web VM
4. Close all services at Peer1 (by end of year)
5. Buy a dumb storage server (~$2000 for 8 TB, or 4 TB raid mirror) for
offsite backups to be hosted by Michael Smith in NH where the WMS
shootout servers live. Migrate the backup vm there, wholesale.
6. Move the wiki to OSGeo logins and https logins
7. Get all projects with stuff on osgeo to declare what they need backed
up, and ask them to create cron type scripts we can execute with Bacula.
I'm available to help, specially with the:
"6. Move the wiki to OSGeo logins and https logins"
Maybe Martin can add more details on this issue.
Meanwhile, I'm testing in our Local Chapter such integration (LDAP + Drupal + MediaWiki + Wordpress).
Related to "3. Migrate Drupal to OSUOSL Web VM", I'm also available to help to go from Drupal 5.x to Drupal 7.x. Drupal 7.x has more sophisticated modules, to work with:
- forms (and also "commerce" to improve our "Donate" button, and to sell OSGeo mugs)
- LDAP
- contents (we can create more contents types, and get more people involved in the web site, to better mirror a live community)
Regards,
Jorge Gustavo
On 20-09-2011 07:34, Alex Mandel wrote:
Since Martin wanted more details, here's the grand scheme for the next
few months. I'll probably try to call a meeting in October to work out
the details.
1. Put up a new download server at OSUOSL
2. Migrate trac to 0.12
3. Migrate Drupal to OSUOSL Web VM
4. Close all services at Peer1 (by end of year)
5. Buy a dumb storage server (~$2000 for 8 TB, or 4 TB raid mirror) for
offsite backups to be hosted by Michael Smith in NH where the WMS
shootout servers live. Migrate the backup vm there, wholesale.
6. Move the wiki to OSGeo logins and https logins
7. Get all projects with stuff on osgeo to declare what they need backed
up, and ask them to create cron type scripts we can execute with Bacula.
On Mon, Sep 19, 2011 at 11:34:59PM -0700, Alex Mandel wrote:
Since Martin wanted more details, [...]
I'm probably the only Primary Admin who was unable to attend FOSS4G,
therefore I'm happy to learn what has been discussed and planned -
especially in those areas where _I_ do feel a certain reponsibility
7. Get all projects with stuff on osgeo to declare what they need backed
up, and ask them to create cron type scripts we can execute with Bacula.
I think it's entirely sufficient to ask OSGeo infrastructure users to
stick to a reasonable directory schema and, even though I appreciate
custom solutions where appropriate, I think the schema which is laid
out by nowadays Linux Distributions in general (and Debian in
particular) serves not only as a pretty good start but maybe even as a
viable compromise.
More details later - cheers,
Martin.
--
Unix _IS_ user friendly - it's just selective about who its friends are !
--------------------------------------------------------------------------
On Mon, Sep 19, 2011 at 11:34:59PM -0700, Alex Mandel wrote:
6. Move the wiki to OSGeo logins and https logins
Enforcing SSL encryption on HTTP login pages is an item I'd like to
enable rather sooner than later. Check it out on 'wiki2.osgeo.org', my
LDAP authentication testbed - I'd be happy to port the change over to
the 'regular' wiki, if nobody objects.
Cheers,
Martin.
--
Unix _IS_ user friendly - it's just selective about who its friends are !
--------------------------------------------------------------------------
On Tue, Sep 20, 2011 at 11:59:53AM +0100, Jorge Gustavo wrote:
Related to "3. Migrate Drupal to OSUOSL Web VM", I'm also available to
help to go from Drupal 5.x to Drupal 7.x. Drupal 7.x has more
sophisticated modules, to work with:
[...]
- LDAP
Does Drupal 7.x require any additional LDAP schemas or does it work
right out of the box using the well-known user/group schema we're using
right now ?
Cheers,
Martin.
--
Unix _IS_ user friendly - it's just selective about who its friends are !
--------------------------------------------------------------------------
On Tue, Sep 20, 2011 at 11:59:53AM +0100, Jorge Gustavo wrote:
Related to “3. Migrate Drupal to OSUOSL Web VM”, I’m also available
to help to go from Drupal 5.x to Drupal 7.x. Drupal 7.x has more
sophisticated modules, to work with:
[…]
LDAP
Does Drupal 7.x require any additional LDAP schemas or does it work
right out of the box using the well-known user/group schema we’re using
right now ?
I have drupal 7 running but haven’t been able to get ldap to work. Some do report that it does though. I don’t think it needs anything special.
On Tue, Sep 20, 2011 at 2:21 PM, Martin Spott <Martin.Spott@mgras.net> wrote:
On Mon, Sep 19, 2011 at 11:34:59PM -0700, Alex Mandel wrote:
6. Move the wiki to OSGeo logins and https logins
Enforcing SSL encryption on HTTP login pages is an item I'd like to
enable rather sooner than later. Check it out on 'wiki2.osgeo.org', my
LDAP authentication testbed - I'd be happy to port the change over to
the 'regular' wiki, if nobody objects.
Please - this will hopefully stop the spam hell in the Wiki.
I am willing to help with the user page migration if needed.
On Tue, Sep 20, 2011 at 02:21:59PM +0200, Martin Spott wrote:
On Mon, Sep 19, 2011 at 11:34:59PM -0700, Alex Mandel wrote:
> 6. Move the wiki to OSGeo logins and https logins
Enforcing SSL encryption on HTTP login pages is an item I'd like to
enable rather sooner than later. Check it out on 'wiki2.osgeo.org', my
LDAP authentication testbed - I'd be happy to port the change over to
the 'regular' wiki, if nobody objects.
Done,
Martin.
--
Unix _IS_ user friendly - it's just selective about who its friends are !
--------------------------------------------------------------------------
On Mon, Sep 19, 2011 at 11:34:59PM -0700, Alex Mandel wrote:
Since Martin wanted more details, [...]
"More details" !? Don't you think it's a valid, _natural_ expectation
from a Primary Admin to get an idea about what has been negotiated in
Denver ?
[...] here's the grand scheme for the next
few months. I'll probably try to call a meeting in October to work out
the details.
[...]
Add your item here...
8. Introduce a culture of shaping decision making processes into a more
openly style.
Cheers,
Martin.
--
Unix _IS_ user friendly - it's just selective about who its friends are !
--------------------------------------------------------------------------
Hi,
as far as I know we did not really negotiate anything much in Denver.
Maybe we should have taken the opportunity to do more but the most
important task we did was to ask Alex to chair SAC. Apart from that we
talked about linking Wiki with LDAP - and that project seems to be under
way now with Jorge, Martin and Alex. We spent five minutes talking about
the membership application and I have the task to flesh out the
corresponding Wiki page to get something small enough so we can get it
done.
Anything else from Denver?
Me not being anywhere near as into SAC as most others on this list can
only guess but I did get the impression that we have an issue with our
backup system. As this is a core service of OSGeo to it's projects I
would like to better understand where we are now and what we might need
to do to do better.
And I can also still feel some frustration and friction and would like
to mediate us through this so that we can proceed with good
collaboration. Let me know if you think I can be of any help.
Thanks,
Arnulf.
Hi Alex,
On Mon, Sep 19, 2011 at 11:34:59PM -0700, Alex Mandel wrote:
Since Martin wanted more details, [...]
"More details" !? Don't you think it's a valid, _natural_ expectation
from a Primary Admin to get an idea about what has been negotiated in
Denver ?
[...] here's the grand scheme for the next
few months. I'll probably try to call a meeting in October to work out
the details.
[...]
Add your item here...
8. Introduce a culture of shaping decision making processes into a more
openly style.
Items 1-6 were all discussed in some way in Denver in bits and pieces.
Some of which were urgent, others of which the relevant admins happened to be in the room.
Item 7 was just another kickstart to revisiting the topic of ensuring we have proper backups of important stuff, but was in no way decided on the best approach.
So to recap:
1. Put up a new download server at OSUOSL - done (Frank)
2. Migrate trac to 0.12 - in progress (Chris)
3. Migrate Drupal to OSUOSL Web VM - needs a champion
4. Close all services at Peer1 (by end of year) - same as #3
5. Buy a dumb storage server (~$2000 for 8 TB, or 4 TB raid mirror) for
offsite backups to be hosted by Michael Smith in NH where the WMS
shootout servers live. Migrate the backup vm there, wholesale.
- Something to discuss, and then send to the board for approval. Technically all the right people have agreed that this plan is feasible
6. Move the wiki to OSGeo logins and https logins - (Martin)
7. Backups - my biggest concern is actually the various databases used for some of the websites (mysql, postgres etc) and ensuring they get dumped and included in backups. Flat file backup no matter the folder structure seems a simpler task - if we define what goes where, we can just ask projects to follow that. Please write up suggestions on a wiki page so we can compare the options and hopefully reach a consensus on approach.
Thanks,
Alex
On 09/29/2011 09:08 AM, Seven (aka Arnulf) wrote:
Hi,
as far as I know we did not really negotiate anything much in Denver.
Maybe we should have taken the opportunity to do more but the most
important task we did was to ask Alex to chair SAC. Apart from that we
talked about linking Wiki with LDAP - and that project seems to be under
way now with Jorge, Martin and Alex. We spent five minutes talking about
the membership application and I have the task to flesh out the
corresponding Wiki page to get something small enough so we can get it
done.
Anything else from Denver?
Me not being anywhere near as into SAC as most others on this list can
only guess but I did get the impression that we have an issue with our
backup system. As this is a core service of OSGeo to it's projects I
would like to better understand where we are now and what we might need
to do to do better.
And I can also still feel some frustration and friction and would like
to mediate us through this so that we can proceed with good
collaboration. Let me know if you think I can be of any help.
Thanks,
Arnulf.
Hi Alex,
On Mon, Sep 19, 2011 at 11:34:59PM -0700, Alex Mandel wrote:
Since Martin wanted more details, [...]
"More details" !? Don't you think it's a valid, _natural_ expectation
from a Primary Admin to get an idea about what has been negotiated in
Denver ?
[...] here's the grand scheme for the next
few months. I'll probably try to call a meeting in October to work out
the details.
[...]
Add your item here...
8. Introduce a culture of shaping decision making processes into a more
openly style.
3. Migrate Drupal to OSUOSL Web VM - needs a champion
4. Close all services at Peer1 (by end of year) - same as #3
3. is on my list for trying to wrap up before mid Nov. and underway
4. I won't touch mailman, but will tackle conference, journal systems
and foss4g websites.
I need help getting https running on the web vm before I can complete
the drupal migration. Anyone can help me with that?
On Thu, Sep 29, 2011 at 11:48 PM, Alex Mandel
<tech_dev@wildintellect.com> wrote:
Items 1-6 were all discussed in some way in Denver in bits and pieces.
Some of which were urgent, others of which the relevant admins happened to
be in the room.
Please post the minutes from Denver.
...
5. Buy a dumb storage server (~$2000 for 8 TB, or 4 TB raid mirror) for
offsite backups to be hosted by Michael Smith in NH where the WMS
shootout servers live. Migrate the backup vm there, wholesale.
I am not much convinced that putting the backup device physically
distant will work due to the large amount of data and network latency.
- Something to discuss, and then send to the board for approval. Technically
all the right people have agreed that this plan is feasible
On Thu, Sep 29, 2011 at 11:48 PM, Alex Mandel
<tech_dev@wildintellect.com> wrote:
Items 1-6 were all discussed in some way in Denver in bits and pieces.
Some of which were urgent, others of which the relevant admins happened to
be in the room.
Please post the minutes from Denver.
There was no meeting, so there are no minutes. I walked around the code
sprint and talked to 1-3 people at a time when they came up for air.
The todo list was all the ideas collectively written down. We do have
the intention of holding a meeting in October on IRC - I'll take time
suggestions that work for everyone.
...
5. Buy a dumb storage server (~$2000 for 8 TB, or 4 TB raid mirror) for
offsite backups to be hosted by Michael Smith in NH where the WMS
shootout servers live. Migrate the backup vm there, wholesale.
I am not much convinced that putting the backup device physically
distant will work due to the large amount of data and network latency.
That's a good question. We should probably refine that by checking how
much data actually gets moved on a nightly basis. Offsite backup is
something of concern, as you may have noted with the recent power issues
we had in San Diego. I'll note the nightly rsync of the download server
to the backup server from San Diego to Oregon has not been a problem.
- Something to discuss, and then send to the board for approval. Technically
all the right people have agreed that this plan is feasible
Really? Who was that?
Frank who runs the rsync backup of the download -> backup server and
Michael who has the physical rack it would go into (The idea can only
work if someone has a rack for such a machine). I realize now we weren't
thinking about Bacula at that point, so we'll need to see what Martin
thinks. Sure there are still details to work out, and that will be a big
topic at our next meeting TBD. Several others nodded that the idea of a
cheap large storage has been on the list of ideas since we bought the 2
machines at OSUOSL.
On Fri, Sep 30, 2011 at 11:45:28AM -0700, Alex Mandel wrote:
That's a good question. We should probably refine that by checking how
much data actually gets moved on a nightly basis. Offsite backup is
something of concern, as you may have noted with the recent power issues
we had in San Diego.
Breakages like the power outages in San Diego are, statistically,
almost as unlikely as a disconnected internet uplink at whichever
hosting site. Just to put a random example. I wouldn't consider this
as a substantial argument.
An incomplete 'full' backup, limited by current disk space on the
backup VM, is, currently, approx. 110 GByte. That's just the bare
minimum and still requires a lot of work to get one ore more machines
back into a working state,
Martin.
--
Unix _IS_ user friendly - it's just selective about who its friends are !
--------------------------------------------------------------------------
That's a good question. We should probably refine that by checking how
much data actually gets moved on a nightly basis. Offsite backup is
something of concern, as you may have noted with the recent power issues
we had in San Diego. I'll note the nightly rsync of the download server
to the backup server from San Diego to Oregon has not been a problem.
Anyone shared their thoughts on setting up mirrors? I suspect that some of
our academic friends (e.g. Uni Nottingham and SIGTE - both running
open source geo labls) might be interested in doing this. I'll reach out and
see.
On Fri, Sep 30, 2011 at 11:45:28AM -0700, Alex Mandel wrote:
[...] I realize now we weren't
thinking about Bacula at that point, so we'll need to see what Martin
thinks.
Have an LTO tape library,
Martin.
--
Unix _IS_ user friendly - it's just selective about who its friends are !
--------------------------------------------------------------------------
On Fri, Sep 30, 2011 at 12:37 PM, Tyler Mitchell <tmitchell@osgeo.org> wrote:
Anyone shared their thoughts on setting up mirrors? I suspect that some of
our academic friends (e.g. Uni Nottingham and SIGTE - both running
open source geo labls) might be interested in doing this. I'll reach out and
see.
Tyler,
I'm not too worried about download.osgeo.org mirrors. It's
not that large. I'm mirroring it on a personal server in
Montreal now. Eventually we might mirror on a server at
US ACE and/or in a VM at telascience.
But, should someone want to mirror it the rsync instructions at:
Best regards,
--
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up | Frank Warmerdam, warmerdam@pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush | Geospatial Software Developer
On Fri, Sep 30, 2011 at 12:37 PM, Tyler Mitchell <tmitchell@osgeo.org> wrote:
Anyone shared their thoughts on setting up mirrors? I suspect that some of
our academic friends (e.g. Uni Nottingham and SIGTE - both running
open source geo labls) might be interested in doing this. I'll reach out and
see.
Tyler,
I'm not too worried about download.osgeo.org mirrors. It's
not that large. I'm mirroring it on a personal server in
Montreal now. Eventually we might mirror on a server at
US ACE and/or in a VM at telascience.
But, should someone want to mirror it the rsync instructions at:
