Re: [Spacewalk-devel] Announcing Spacewalk 0.3
Vladimir Zlatkin wrote: Clifford Perry wrote: Vladimir Zlatkin wrote: Jesus M. Rodriguez wrote: * rhn-satellite is no longer in /etc/init.d, use /sbin/rhn-satellite to start/stop the entire satellite. I am curious, what is the motivation for this change? Various levels of breakage in having on chkconfig style script owning multiple daemons. Move to a more per daemon init script and have /sbin/ command to stop/start them all at once if needed once booted up. Allow for the daemons to start cleanly also when switching run levels. Michael Mraka had 3 bugs and so tackled the issue by re-factoring this structure. A suggestion (which I hope made it into bugzilla) was made last week to put a simple echo please use the /sbin/ .. command now when attempting to run /etc/init.d/rhn-satellite. Do you see something 'bad' in this? Or just wondering about why? I don't see it as a bad change, if bugs were fixed then this is a good thing. Monitoring scripts and clustering applications will have to change. Also, /sbin is not a typical location for init scripts. That is why I was curious. Why can't this can't be dealt with by multiple-init scripts with dependencies? I don't like the idea of init.d scripts that don't have a function being installed by an RPM either. --Michael ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel
Re: [Spacewalk-devel] root as contributor
Mike McCune wrote: Miroslav Suchý wrote: Welcome root :) http://miroslav.suchy.cz/spacewalk/gitstat/commit-detail.php?commit=6cdf4423e04ffd66fda796e4946a662a9a03e497 and 3 others commits Can you please use your real identity for commits? Thx. Anyone have any good tips on how we can change the author on those commits? I asked this question previously, with a different reason in mind -- attributing patches that didn't get committed with -m (which I normally don't do). Short answer: git-commit with --ammend can be used to fix the /last/ commit Want to do more? Generally it's considered dangerous to muck with the history. (That's why it's called history) Did find some things on Google but I wouldn't recommend it: http://blog.jacius.info/articles/2008/6/22/git-tip-fix-a-mistake-in-a-previous-commit Basically if you can avoid editing code as root, it becomes easier :) --Michael ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel
Re: [Spacewalk-devel] changing of progress indicator?
Jan Pazdziora wrote: On Thu, Nov 13, 2008 at 09:06:06PM -0500, Jesus M. Rodriguez wrote: Jan, Um really? :) Were the lines boring you? We needed something to visually distinguish the new version from old versions right at the beginning, in the installer. print I'm the new version! ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel
Re: [Spacewalk-devel] Spacewalk and SELinux: progress status
I don't have selinux on. Is this expected? it took almost 10 minutes and I eventually just CTRL+C-ed it. Well yes, we run restorecon on /var/satellite to set correct context, even if you are not in Enforcing. It is not expected to fail thou. Also calling restorecon with selinux disabled probably won't work. Cobbler makes sure it doesn't bother with restorecon in those instances. Enable in /etc/selinux, touch /.autorelabel, and reboot? You'd want to strive for 100% clean runs in setroubleshoot. FYI -- Recently Dan Walsh recommended I /not/ support SELinux on EL 4, and only do EL 5 since it was more manageable (supports public_content_t).Seeing this impacts spacewalk, I would suggest spacewalk take the same position. --Michael ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel
Re: [Spacewalk-devel] #!/usr/bin/env python to #!/usr/bin/python
Clifford Perry wrote: Jan Pazdziora wrote: On Wed, Jan 14, 2009 at 08:57:38AM -0400, Devan Goodwin wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Actually, this AVC denial is about different problem. So it was not a good example. But nevertheless: shouldn't we decide for env or direct path in the shebang line and be consistent about it? +1. I'd go the other direction though and stick with /usr/bin/env python, iirc that's considered best practice to accommodate people with Python installed in a weird location or using multiple versions. If you have Python in a weird location, you probably won't have osa-dispatcher .py files installed in its PYTHONHOME, will you? So, the first import which assumes that the python you run actually has all the prerequisites installed, will fail. Alternatively, mixing different pythons and libraries from different pythons might produce weird results because symbols referenced in one library might not be present in the library from that second python. We actually have (non-public) bugzilla about this very problem. I'd argue that we should stop pretending that /usr/bin/env python will work in the general case, any just put /usr/bin/python there. If someone needs to run it with different interpreter, they can always do python /the/path/to/the/script I would prefer the hard code path to the python binary for reasons stated by Jan above. Folks - other than preference normally - please give feedback based on this above information. Cliff I'd suggest /usr/bin/python without the env. If the RPM doesn't work with the distribution-specific Python, things are quite busted, and folks shouldn't expect them to work. We do not package two python versions, and the modules won't be properly installed as required by the RPM. --Michael ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel
Re: [Spacewalk-devel] Informal Devel Environment Survey
Jesus M. Rodriguez wrote: On Thu, Jan 22, 2009 at 3:03 PM, Coe, Colin C. (Unix Engineer) colin@woodside.com.au wrote: I've found that doing the steps under 'Deploying Development Schema' doesn't work (for me anyway) and ends up needing to redo the dev environment. Also, I'd like to see https://fedorahosted.org/spacewalk/wiki/JavaDesign fleshed out a lot more. Anything in particular? I'd be happy to update it. jesus ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel Could the dev-environment be more appliancey? How about a shell-script/recipe to automate environment setup, or publishing a kickstart for installation of a dev-environment in a virtual machine (with just the virt-install command and kickstart, you should be good to go)? One problem is grabbing the Oracle bits, for now, so that may have to be a one-off, but everything else, perhaps... --Michael ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel
Re: [Spacewalk-devel] Informal Devel Environment Survey
Jesus M. Rodriguez wrote: On Fri, Jan 23, 2009 at 12:27 PM, Michael DeHaan mdeh...@redhat.com wrote: Jesus M. Rodriguez wrote: On Thu, Jan 22, 2009 at 3:03 PM, Coe, Colin C. (Unix Engineer) colin@woodside.com.au wrote: I've found that doing the steps under 'Deploying Development Schema' doesn't work (for me anyway) and ends up needing to redo the dev environment. Also, I'd like to see https://fedorahosted.org/spacewalk/wiki/JavaDesign fleshed out a lot more. Anything in particular? I'd be happy to update it. jesus ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel Could the dev-environment be more appliancey? How about a shell-script/recipe to automate environment setup, or publishing a kickstart for installation of a dev-environment in a virtual machine (with just the virt-install command and kickstart, you should be good to go)? One problem is grabbing the Oracle bits, for now, so that may have to be a one-off, but everything else, perhaps... The appliance idea is a decent one, and worth adding to the list of dev setups. I personally use a virt guest to do my development in. I wouldn't want the appliance to be the only way of dev setup. A great idea though. jesus ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel By appliance I mean a virt guest with a kickstart and and embedded/recipe script that gets everything working. It should not be an image. --Michael ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel
Re: [Spacewalk-devel] PATCH: Add web UI for managing Cobbler Snippets
Coe, Colin C. (Unix Engineer) wrote: Hi All Attached is a patch that adds web UI for managing Cobbler Snippets. The only problem with this that I am aware of is the inability to delete snippets that have been created, however I believe that this is due to the tomcat5 policy. Also, many thanks to those on the mailing list and on #spacewalk-devel that helped with this work. Comments/criticism welcome. Thanks CC This is a good idea for Satellite, though ideally I'd like to see Cobbler Web have the same feature -- it's just as useful there. Adding/deleting, and enumerating of cobbler snippets should /ideally/ happen over the Cobbler API to also make this more scriptable. In this case I'd like to see associated changes made in Cobbler's remote.py and it done that way, rather than assuming too much about cobbler's filesystem layout and storage, which may be subject to change. --Michael NOTICE: This email and any attachments are confidential. They may contain legally privileged information or copyright material. You must not read, copy, use or disclose them without authorisation. If you are not an intended recipient, please contact us at once by return email and then delete both messages and all attachments. ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel
[Spacewalk-devel] Cobbler architecture diagram
I'm putting together some more visual information on Cobbler, though I thought spacewalk folks would find this useful. https://fedorahosted.org/cobbler/attachment/wiki/ChartsAndGraphs/code.png More pictures to come. --Michael ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel
Re: [Spacewalk-devel] Cobbler architecture diagram
Michael DeHaan wrote: I'm putting together some more visual information on Cobbler, though I thought spacewalk folks would find this useful. https://fedorahosted.org/cobbler/attachment/wiki/ChartsAndGraphs/code.png More pictures to come. --Michael ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel Added a bit more, https://fedorahosted.org/cobbler/wiki/ChartsAndGraphs There's a object graph picture and also a diagram that shows what cobbler replication can do. --Michael ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel
Re: [Spacewalk-devel] Dojo?
Cliff wrote: So, anyone looked at dojo in the pass? If so, thoughts on it? Could/would it replace our list tag? Make it better? Compatible licenses? The old age problem of how to represent lots of data in a timely manner to Spacewalk folks and allow them to sort/select/filter/paginate that data. Not saying 'we need to be using this!' - but more interested in what others think about it :) http://en.wikipedia.org/wiki/Dojo_Toolkit Thanks, Cliff ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel Sidenote, no Fedora packaging guidelines for javascript libraries yet -- though one problem we face is that all the compressors are non-free, thus we can't include the compressed versions in any RPMs. Personally I think this is a huge roadblock Fedora needs to solve somehow. Until then, apps could include such libraries but cannot package the compressed versions IIRC. IMHO, javascript is a requirement for any modern web application, and those that use things like NoScript would need to grant exemptions. --Michael ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel
[Spacewalk-devel] FYI -- Cobbler 2.0
Cobbler 2.0 is now available, and should be in Fedora/EPEL testing very soon. https://fedorahosted.org/pipermail/cobbler/2009-September/004983.html Testing w/ Spacewalk welcome. --Michael ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel
Re: [Spacewalk-devel] Cobbler breaking re-provisioning?
On 10/07/2009 11:51 AM, Duncan Innes wrote: Hi, I'm testing this on Satellite 5.3 at the moment, but the inclusion of Cobbler appears to break re-provisioning. My kickstarts use a postinstall script to configure the network scripts according to which configuration files are installed by the relevant activation keys. Unfortunatley, cobbler seems to write it's own network config files and then deletes everything that may have existed proir to that. So I can't even push anything to /etc/sysconfig/network-scripts. The erroneous part of cobbler.ks is as follows: %post ( /opt/edft/bin/postinstall ) /root/ks-post.log 21 # Start post_install_network_config generated code mkdir /etc/sysconfig/network-scripts/cobbler cp /etc/sysconfig/network-scripts/ifcfg-lo /etc/sysconfig/network-scripts/cobbler/ # Start configuration for eth0 IFNAME=$(ifconfig -a | grep -i '00:24:81:E3:F9:78' | cut -d ' ' -f 1) if [ -f /etc/modprobe.conf ] [ $IFNAME ]; then grep $IFNAME /etc/modprobe.conf | sed s/$IFNAME/eth0/ /etc/modprobe.conf.cobbler grep -v $IFNAME /etc/modprobe.conf /etc/modprobe.conf.new rm -f /etc/modprobe.conf mv /etc/modprobe.conf.new /etc/modprobe.conf fi echo DEVICE=eth0 /etc/sysconfig/network-scripts/cobbler/ifcfg-eth0 echo HWADDR=00:24:81:E3:F9:78 /etc/sysconfig/network-scripts/cobbler/ifcfg-eth0 echo ONBOOT=yes /etc/sysconfig/network-scripts/cobbler/ifcfg-eth0 echo BOOTPROTO=dhcp /etc/sysconfig/network-scripts/cobbler/ifcfg-eth0 # End configuration for eth0 rm -f /etc/sysconfig/network-scripts/ifcfg-* mv /etc/sysconfig/network-scripts/cobbler/* /etc/sysconfig/network-scripts/ rm -r /etc/sysconfig/network-scripts/cobbler if [ -f /etc/modprobe.conf ]; then cat /etc/modprobe.conf.cobbler /etc/modprobe.conf rm -f /etc/modprobe.conf.cobbler fi # End post_install_network_config generated code # Start koan environment setup echo export COBBLER_SERVER=vstlbsatx01.edftrading.com /etc/profile.d/cobbler.sh echo setenv COBBLER_SERVER vstlbsatx01.edftrading.com /etc/profile.d/cobbler.csh # End koan environment setup wget http://vstlbsatx01.edftrading.com/cblr/svc/op/ks/system/tcplrdacs01.edftrading.com:1; -O /root/cobbler.ks wget http://vstlbsatx01.edftrading.com/cblr/svc/op/trig/mode/post/system/tcplrdacs01.edftrading.com:1; -O /dev/null Is this happening from Spacewalk too? Duncan ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel Rather than point blame about what you /think/ is wrong, I'd recommend just stating the problem you are seeing and let us go from there. In this case, Cobbler is not broken. Your usage of your kickstart template, however, is incompatible with them. Cobbler network objects eliminate the need to have seperate needs to configuring networking. If you are doing your own thing, remove the network snippets from the Cobbler template. --Michael ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel
Re: [Spacewalk-devel] Cobbler breaking re-provisioning?
I'm not making any accusations here - I'm only asking if this was decided to be the new behaviour for Spacewalk/Satellite/Cobbler. Should there be an inconsistency between fresh kickstart installs and re-provisions? It is intended that, if you have the network configuration snippets referenced in your kickstart templates, that you will be using the network configuration scripts that Cobbler generates rather than pushing out your own. If you don't have those snippets there, things just work. If existing Spacewalk generated (templated, not raw uploaded) scripts insert these snippets blindly into existing kickstarts, that might be a bug in Satellite, as they should generate the same kickstart output (essentially) both before and after the upgrade -- I agree with that. I don't agree that Cobbler broke it... Spacewalk's particular usage of Cobbler, yes. If you are using self-uploaded-into-Spacewalk kickstart templates based on /var/lib/cobbler/kickstarts/sample*.ks, then you would need to remove the network configuration portions if you wanted to write your own things in %post. I can mostly speak for Cobbler and not Spacewalk, so perhaps Spacewalk folks can weigh in with what you'd need to do. Thanks Duncan Innes ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel
