I'm a Linux administrator since 1999, actually I'm a RHCE working for
an Red Hat Partner in Venezuela (http://www.sldv777.com), I'm a
volunteer on others Fedora Project, I have an interest in helping out
with the Fedora Infrastructure Project, I've done some scripting
mainly Bash and Perl, i know a lot of stuff on areas such as
Networking, Monitoring, Email, FDS/LDAP.
Let me know guys how to start, i has requested membership in the
Wilmer Jaramillo M.
GPG Key Fingerprint = 0666 D0D3 24CE 8935 9C24 BBF1 87DD BEA2 A4B2 1E8A
At about 11pm CDT, I noticed db1 stopped responding to pgsql database
connections to either the mirrormanager database or the FAS database.
One can start to connect to it:
psql -h db1 -d mirrormanager -U mirroradmin -W
Password for user mirroradmin:
but it just hangs at this point. Other apps trying to access it are
I don't have access to db1, nor do I know enough about pgsql to feel
comfortable digging. Hopefully someone who has both access and skill
can take a look?
Dell Linux Solutions linux.dell.com & www.dell.com/linux
Linux on Dell mailing lists @ http://lists.us.dell.com
This morning at some point the wiki went down again. Looking at the logs on
the test2 xen guest it looks like this was the last thing accessed:
10.8.32.56 - - [15/Apr/2007:04:03:39 -0700] "GET /wiki/Releases/7 HTTP/1.1"
200 14870 "http://fedoraproject.org/wiki/Releases/7/Schedule" "Mozilla/5.0
(X11; U; Linux i686; en-US; rv:18.104.22.168) Gecko/20070410 Fedora/22.214.171.124-3.fc7
There were some http processes still going but in the error log there was:
[Sun Apr 15 05:16:48 2007] [notice] child pid 3299 exit signal Segmentation
[Sun Apr 15 05:16:50 2007] [notice] child pid 3301 exit signal Segmentation
[Sun Apr 15 05:16:50 2007] [notice] child pid 3302 exit signal Segmentation
[Sun Apr 15 05:16:51 2007] [notice] child pid 3303 exit signal Segmentation
a bunch of those. I restarted http and the wiki came right back up. This was
at Sun Apr 15 05:17:02 MST 2007 according to the box, so roughly an hour
I don't know if the other outages play a role in this crash or not, as it
seemed like we had a good config going.
Release Engineer: Fedora
How difficult would it be to create for Fedora something like this:
(not for downloads but for installs)
Jabber ID: glezos(a)jabber.org, GPG: 0xA5A04C3B
"He who gives up functionality for ease of use
loses both and deserves neither." (Anonymous)
The wiki is having an outage, we're working on it. Should be up soon.
We suspect our issues are because of mod_python. The moin developers
suggest not using it. So we'll go back to cgi (slow) and deal with it
Peter van der Does will be conducting a DNS security audit next week;
please speak up if this is going to be a problem for anyone.
ns1.fedoraproject.org : Tuesday - 10 AM EST
dns1.j2solutions.net : Wednesday - 10 AM EST
The results will eventually make their way onto the InfrastructurePrivate wiki.
For internal outages I've created two scripts on each of the proxy servers.
They're pretty self explanatory. Basically though they stop httpd and
start up a special instance. Configs, html and images are all self
contained in /etc/httpd/outage/ so as long as that dir exists and apache
is installed, the outage page should be viewable.
My name is Frank and I would like to volunteer to help.
I've already traded a few emails with Mike McGrath. He suggested that I send
a short introduction the list. So here it is.
I am 58 years old. I have been writing software for about 37 years
professionally - longer if you count college. But of course when I was in
college, we had to suffer with punched cards and paper tape.
I have been doing System Administration on and off the whole time usually in
combination with some application development work.
I am currently working on SGI (IRIX) hardware. But have also worked on Sun
(Solaris and Sun/OS), VAXes and Control Data (CDC).
I have written lots of Perl scripts and shell scripts. I have used several
different shells over the years. Most of my recent scripts have been written
in Korn shell.
I have not used Python to date. I am reading "Learning Python" online. This is
thanks to my employer who has granted me access to Safari.
I have some knowledge of databases but am not an expert by any means. I can
write some basic SQL.
I don't have anything specific in mind for an area in which to help. So I'm
If you have any questions, please ask away.
If you want my help, please speak up.
Currently, the FAS DB has a field for domains::
CREATE TABLE role (
person_id INTEGER NOT NULL REFERENCES person (id),
project_group_id INTEGER NOT NULL REFERENCES project_group (id),
--- role_domain would be relevant for limiting the person to a
specific portion of the project,
--- e.g. '/rpms/util-linux'
UNIQUE (person_id, project_group_id, role_domain)
The intention of this definition is for a person to belong to a group
within a domain no more than once. However, it doesn't work out that
way. The value of role_domain is allowed to be null in the above
definition. Among SQL databases, Null is the absence of a value and
postgres (among other databases) takes this to mean that a Null value
has no effect on the unique clause. So it's legal to have several
entries for the same person_id and project_group_id as long as the
role_domain is null. Additionally, a role_domain of "" (empty string)
and Null are different values as well.
These issues are causing us to see some users having duplicate groups
defined in the FAS. In order to fix this, I would like to change the
database schema to make role_domain non-nullable. Concurrently, I'll be
converting all roles where a group is defined as null into a group where
the group is defined as "" and merging any that have been duplicated.
This fix also requires auditing the FAS1 code for places where
role_domain is defined as Null and changing them to "".
If no one objects, I'll try to get this done by Friday (although it's
turning into a busy week at work so I may have to push this off until
Monday. I'll be sure to inform the list when the changes go live.)