From Mageia wiki
Jump to: navigation, search
(Contact/join the team)
(Useful pages: remove "Test")
(31 intermediate revisions by 7 users not shown)
Line 1: Line 1:
 +
[[File:Valstar.jpg|300px|right]]
 
= Contact/join the team =
 
= Contact/join the team =
* join the [irc://irc.freenode.net/#mageia-sysadm] IRC channel and/or subscribe to [https://mageia.org/mailman/listinfo/mageia-sysadm mageia-sysadm] mailing-list
+
* join the [irc://irc.freenode.net/#mageia-sysadm IRC channel] and/or subscribe to the [https://ml.mageia.org/l/info/sysadmin-discuss sysadmin-discuss] mailing-list
* team is built from: blino, buchan, colin, dams, dmorgan, misc, nanardon, pterjan, tmb
+
* team is built from: see [http://people.mageia.org/g/mga-sysadmin.html mga-sysadmin ldap group]
* team leader & [[org:council] representative: ?
 
 
 
Team weekly meetings happen on (which day?) at (what time?) UTC (todo, link to logs/notes archives)
 
  
 
= Participating to the team =
 
= Participating to the team =
The whole infrastructure is managed by Puppet and SVN. The repository is public, and can be found at: http://svnweb.mageia.org/adm/
+
The whole infrastructure is managed by Puppet and git. The repository is public, and can be found at http://gitweb.mageia.org/ under the Infrastructure section.
  
If you wish to join the team, the recommended way is to study how the infrastructure is organized, and start sending patchs, like any free software project.
+
If you wish to join the team, the recommended way is to study how the infrastructure is organized, and start sending patches as in any free software project.
  
 
To get a checkout of our puppet configuration, you use :
 
To get a checkout of our puppet configuration, you use :
   svn co svn://svn.mageia.org/svn/adm/puppet/
+
   git clone git://git.mageia.org/infrastructure/puppet
  
To test it, you should use a virtual machine, running Mandriva 2010.1, and you will need a recent puppet backport ( > 2.6.0 ). Any virtualisation technology should work.  
+
To test it, you should use a virtual machine, running Mageia, and you will need a recent puppet backport ( > 2.6.0 ). Any virtualisation technology should work.  
  
 
== Infrastructure clone ==
 
== Infrastructure clone ==
  
 
To replicate our own infrastructure for test purpose, the main puppet server should be on a computer who can be contacted on puppetmaster.$yourdomain.
 
To replicate our own infrastructure for test purpose, the main puppet server should be on a computer who can be contacted on puppetmaster.$yourdomain.
It is much easier to call the computers like we do ( valstar being the main node ), see [http://svnweb.mageia.org/adm/puppet/manifests/nodes.pp?view=markup|manifest/nodes.pp] ). Using our configuration will automatically run a svn checkout in /etc/puppet  
+
It is much easier to call the computers like we do ( valstar being the main node ), see [http://gitweb.mageia.org/infrastructure/puppet/tree/manifests/nodes.pp] ). Using our configuration will automatically checkout in /etc/puppet  
on the puppetmaster, so take care if you plan to use git-svn or similar. For the more advanced tests, a ldap directory  
+
on the puppetmaster, so please be aware of this. For the more advanced tests, a ldap directory  
is needed, even if almost empty.  
+
is needed, even if almost empty.
 
 
== Ldap cloning ==
 
 
 
TODO write this
 
 
 
= Contacting the team =
 
 
 
Most sysadmin can be found on irc, on the channel #mageia-sysadm on irc.freenode.net
 
 
 
You can also subscribe to the discussion list, on https://www.mageia.org/mailman/listinfo/mageia-sysadm , or read the archives on https://www.mageia.org/pipermail/mageia-sysadm/
 
 
 
== Sysadmin tasks ==
 
 
 
See [http://svnweb.mageia.org/soft/build_system]
 
 
 
List of bugs and things to do is now on bugzilla :
 
https://bugs.mageia.org/showdependencytree.cgi?id=859&hide_resolved=1
 
 
 
== List of webservice passwords ==
 
Add here each webservice we use, and who has the password :
 
 
 
*  cia.vc: ?
 
*  dailymotion: dams/boklm/ennael/rda/rtp
 
*  dedibox ( rabbit ): misc, ennael
 
*  dlvr.it: dams/boklm/ennael/rda/rtp
 
*  facebook: dams/boklm/ennael/rda/rtp
 
*  flickr: dams/boklm/ennael/rda/rtp
 
*  gandi dns: dams/boklm/ennael/rda/rtp
 
*  gandi vm: dams/boklm/ennael/rda/rtp
 
*  google: dams/boklm/ennael/rda/rtp
 
*  identica: dams/boklm/ennael/rda/rtp
 
*  ovh dns ( bonjourmageia.fr ):  ?
 
*  tumbler account (bonjourmageia): dams/boklm/ennael/rda/rtp
 
*  twitter: dams/boklm/ennael/rda/rtp
 
*  wordpress.com: dams/boklm/ennael/rda/rtp
 
*  yahoo: dams/boklm/ennael/rda/rtp
 
*  youtube: dams/boklm/ennael/rda/rtp
 
 
 
== SQL Databases ==
 
Postgresql for software that support it, MySQL otherwise.
 
= Services =
 
*  blog: mysql (krampouezh)
 
*  bugzilla: postgresql (alamut)
 
*  epoll: postgresql (alamut)
 
*  mga::mirror: postgresql (alamut)
 
*  package db: mysql (alamut)
 
*  sympa: postgresql (alamut)
 
*  transifex: postgresql (alamut)
 
  
= Servers =
+
= Mailing lists =
*  postgresql on alamut (installed)
 
*  mysql on alamut (TODO)
 
*  mysql on krampouezh (TODO)
 
  
== System Usernames ==
+
* [https://ml.mageia.org/l/info/sysadmin-discuss Sysadmin discussion list]
* iurt : used on build nodes to build packages
+
* Other lists can be found [https://ml.mageia.org/l/lists/sysadmin on this page] (commits, bugs, reports, etc ...)
* mirrorbot : data pushed on the mirror tree automatically by a bot
 
*  mirrorstatic : data pushed on the mirror tree manually and not changing
 
*  schedbot : used on valstar to manage scheduling of package builds
 
*  signbot : used by the script to sign packages
 
  
== Repsys and SVN repository for packages ==
+
= User accounts requests =
*  git branch binrepo-svn for repsys on http://git.mandriva.com/
+
Requests about user accounts or additional permissions should be done using bugzilla on the [https://bugs.mageia.org/enter_bug.cgi?product=Infrastructure&format=guided Infrastructure -> Account request] component.
*  README file for this branch: http://is.gd/gRNHb
 
*  http://kenobi.mandriva.com/~spuk/binrepo.html
 
*  Thread on Maintainers (Mandriva) mailing list: http://lists.mandriva.com/maintainers/2009-08/msg00240.php
 
  
== Building packages for 2010.1 ==
+
= Useful pages =
*  copy your src.rpm on root@jonund.mageia.org:~iurt/2010.1-SRPMS
 
*  run iurt2010.1 as iurt user on jonund to rebuild the package :
 
    ''# su - iurt
 
    $ cd 2010.1-SRPMS
 
    $ iurt2010.1 package.src.rpm''
 
*  run sync2010.1 on valstar as schedbot user :
 
    ''# su - schedbot
 
    $ sync2010.1''
 
*  The package should now be available in /distrib/mandriva/2010.1-mgaserv/x86_64/ and can be installed on the servers with urpmi.
 
  
== LDAP Groups ==
+
* [[Pushing_updates|Pushing package updates]]
Groups for each type of account :
+
* [[SOP_Sysadmin_access_revocation|Removing Sysadm access]]
* packager : mga-packagers and mga-packagers-committers
+
* [[SOP_Certificate_update|Updating SSL Certificates]]
* apprentice packager : mga-packagers-commiters
+
* [[SOP_Version_and_Release_freeze|Controlling version/release freeze]]
* i18n members : mga-i18n
+
* [[SOP_Rename_user|Renaming users]]
* i18n committers : mga-i18n and mga-i18n-committers
 
  
== New hardware needed ==
+
[[Category:Sysadmin]] [[Category:Teams]]
*  new disk to replace failing disk on ecosse
 
*  new disk to replace failing disk on valstar
 
*  2 new disks for build nodes : raid1 for system fs, raid0 for build filesystem for improved performances
 
*  backup server
 
*  ARM build nodes
 
** board suggested by rtp: freescale board http://fr.mouser.com/ProductDetail/Freescale-Semiconductor/MCIMX53-START/?qs=sGAEpiMZZMspbelMIdDJdKswmyqRD7aG also used in http://blog.einval.com/2011/09/05#armhf_buildds
 
*  server to host VMs for packagers
 

Revision as of 22:05, 18 March 2020

Valstar.jpg

Contact/join the team

Participating to the team

The whole infrastructure is managed by Puppet and git. The repository is public, and can be found at http://gitweb.mageia.org/ under the Infrastructure section.

If you wish to join the team, the recommended way is to study how the infrastructure is organized, and start sending patches as in any free software project.

To get a checkout of our puppet configuration, you use :

 git clone git://git.mageia.org/infrastructure/puppet

To test it, you should use a virtual machine, running Mageia, and you will need a recent puppet backport ( > 2.6.0 ). Any virtualisation technology should work.

Infrastructure clone

To replicate our own infrastructure for test purpose, the main puppet server should be on a computer who can be contacted on puppetmaster.$yourdomain. It is much easier to call the computers like we do ( valstar being the main node ), see [1] ). Using our configuration will automatically checkout in /etc/puppet on the puppetmaster, so please be aware of this. For the more advanced tests, a ldap directory is needed, even if almost empty.

Mailing lists

User accounts requests

Requests about user accounts or additional permissions should be done using bugzilla on the Infrastructure -> Account request component.

Useful pages