[Oscar-users] RE : RE : RE : RE :OSCAR on CentOS-6: how to test.(Warning, experimental) WAS:Re: OSCAR on CentOS-6: bootstrap successfull using online repository.
Costel, If si_rmimage can't find the /var/lib/systemimager/images/centos6 image it seems that it was physically removed, but if this image was manually deleted some infos may still be present in oscar database: to check: mysql -u root oscar select * from Images; ^D If you want to remove some entries like centos6 image infos: DELETE FROM Images WHERE name=centos6; Then if some nodes are affected to this image (note the id before deleting it). Let say that centos6 image id was equal to 7 and that centos63 imageid is equal to 8), then to update your nodes definitions: UPDATE Nodes SET image_id=8 WHERE image_id=7; Though, this is hardcore ;-) and not recommanded. IMHO, my recomendation would be to restart from scratch using: oscar-config -t I think that this will remove some problems with post-scripts that failed and left incomplete setups. (like missinc /etc/c3.conf) oscar-config -t will remove the oscar database, the images (hopefully) and some oscar packages. Check that /etc/c3.conf is cleared as well and that /var/lib/systemimager/images/your images are removed. check that mysql -u root oscar fails or do a mysql -u root followed by a drop database oscar; Then after having the correct /etc/oscar/supported_distros.txt (with your local epel) start over with the oscar bootstrapping process and distro setup. I think that you should experience far less problems. Though, there are still some problems. Especially in sync_files that for some unknown reasons is called before nodes are up without the --image-only option resulting in the /etc/c3.conf containing all nodes as dead. My advise: before step 7, check if /etc/c3.conf is present and if yes, look at dead nodes and reenable them is they are tagged as dead. I'm really sorry, I traked this sync_file bug along the day without success so far I've just discovered that c3_conf_mgr which should handle disable/enable for nodes does nothing. Even the help seems wrong: --delnodes whil script looks for --rmnodes options Still digging by adding debug messages that are triggered if $verbose is on. Best regards. -- Olivier LAHAYE CEA DRT/LIST/DCSI/DIR De : Costel Seitan [csei...@slb.com] Date d'envoi : jeudi 28 février 2013 17:09 À : LAHAYE Olivier; oscar-users@lists.sourceforge.net Cc: oscar-de...@lists.sourceforge.net Objet : RE: [Oscar-users] RE : RE : RE :OSCAR on CentOS-6: how to test.(Warning, experimental) WAS:Re: OSCAR on CentOS-6: bootstrap successfull using online repository. Olivier, When trying si_rmimage centos6 I got “Couldn't open /var/lib/systemimager/images/centos6/etc/systemimager/boot/ARCH for reading No such file or directory at /usr/sbin/si_rmimage line 151.” ls /etc/c3.conf ls: cannot access /etc/c3.conf: No such file or directory But does the /etc/c3.conf exists at this step? In fact, adding nodes failed few days ago so I believe OSCAR did not succeed to do more than adding them to /etc/hosts. However, at one moment in time I succeeded to build an image so I will try the same image name. I afterthat wrote it again and again many times without succeeding so maybe the error message above is not a real issue. I created the local repo and I will try to build the image again. Thank you, Costel SEITAN From: LAHAYE Olivier [mailto:olivier.lah...@cea.fr] Sent: Thursday, February 28, 2013 4:11 PM To: Costel Seitan; oscar-users@lists.sourceforge.net Cc: oscar-de...@lists.sourceforge.net Subject: RE : [Oscar-users] RE : RE : RE :OSCAR on CentOS-6: how to test.(Warning, experimental) WAS:Re: OSCAR on CentOS-6: bootstrap successfull using online repository. Hello, Did you removed the old image? If it is still here (in the oda database at list), it's normal that sync_files will try to find the old image. Try to do si_rmimage centos6 BTW, can you tell me if your /etc/c3.conf file has dead nodes (except remove_for_0-indexing which is normal) I found a problem: OSCAR::NodeMgt::update_list_alive_nodes is called somewhere before nodes are up and alive resulting in all nodes marked as dead in /etc/c3.conf resulting in sync_files not syncing anything resulting in oscartst failing the ssh test..domino effect. This routine is called from sync_files, thus I think that sync_files is called before nodes exists (without --image-only) which confirms your previous error. This confirm that fixing one for all sync_files will clear many problems (image failure erroneous messages, c3.conf with dead nodes, files not synced, ) Best regards, Olivier. -- Olivier LAHAYE CEA DRT/LIST/DCSI/DIR De : Costel Seitan [csei...@slb.com] Date d'envoi : jeudi 28 février 2013 15:36 À : LAHAYE Olivier; oscar-users@lists.sourceforge.net Cc: oscar-de...@lists.sourceforge.net Objet : RE: [Oscar-users] RE : RE : RE :OSCAR on CentOS-6: how to test.(Warning, experimental) WAS:Re: OSCAR on CentOS-6: bootstrap successfull using online repository
[Oscar-users] RE : RE : RE : RE :OSCAR on CentOS-6: how to test.(Warning, experimental) WAS:Re: OSCAR on CentOS-6: bootstrap successfull using online repository.
DongInn, I think it's for cependancies like perl-Tk (maybe a few others) Regards, Olivier, -- Olivier LAHAYE CEA DRT/LIST/DCSI/DIR De : DongInn Kim [di...@cs.indiana.edu] Date d'envoi : jeudi 28 février 2013 16:24 À : oscar-users@lists.sourceforge.net Objet : Re: [Oscar-users] RE : RE : RE :OSCAR on CentOS-6: how to test.(Warning, experimental) WAS:Re: OSCAR on CentOS-6: bootstrap successfull using online repository. Costel, I am curious why you need to have the epel repository during the step 4. Regards, -- - DongInn On Feb 28, 2013, at 9:36 AM, Costel Seitan csei...@slb.commailto:csei...@slb.com wrote: Hello, Thank you for your comments Olivier. The nodes should not exist, however I see that the entries were added to /etc/hosts because of my previous tries: .. # OSCAR block: OSCAR hosts 192.168.0.86 epsl86 192.168.0.87 epsl87 192.168.0.88 epsl88 192.168.0.89 epsl89 But, what made me think of rsyncd was the message “rsync: failed to connect to 127.0.0.1: Connection refused (111)” There is something else that I noticed right now. For the last try I decided to name the SIS image “centos63” (compared to “centos6” just before). When analyzing the error messages it looks like some of the commands point to the old image directory (centos6) : ERROR: Impossible to detect the OS for /var/lib/systemimager/images/centos6 at /opt/sync_files/bin/sync_files line 319 main::distro_detect('dir', '/var/lib/systemimager/images/centos6') called at /opt/sync_files/bin/sync_files line 798 ERROR: Impossible to detect the distro (/var/lib/systemimager/images/centos6) at /opt/sync_files/bin/sync_files line 800. ERROR: Impossible to execute /opt/sync_files/bin/sync_files -i centos63 --force --image-only at /var/lib/oscar/packages/sis/api-post-image line 205. I will create the local epel repo and try again. There is no reason to do so but I will eventually name my SIS image “centos6” again. Kind Regards, Costel SEITAN -- Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_feb___ Oscar-users mailing list Oscar-users@lists.sourceforge.netmailto:Oscar-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/oscar-users -- Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_feb___ Oscar-users mailing list Oscar-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/oscar-users
[Oscar-users] RE : RE : RE : RE :OSCAR on CentOS-6: how to test.(Warning, experimental) WAS:Re: OSCAR on CentOS-6: bootstrap successfull using online repository.
Hello, Did you removed the old image? If it is still here (in the oda database at list), it's normal that sync_files will try to find the old image. Try to do si_rmimage centos6 BTW, can you tell me if your /etc/c3.conf file has dead nodes (except remove_for_0-indexing which is normal) I found a problem: OSCAR::NodeMgt::update_list_alive_nodes is called somewhere before nodes are up and alive resulting in all nodes marked as dead in /etc/c3.conf resulting in sync_files not syncing anything resulting in oscartst failing the ssh test..domino effect. This routine is called from sync_files, thus I think that sync_files is called before nodes exists (without --image-only) which confirms your previous error. This confirm that fixing one for all sync_files will clear many problems (image failure erroneous messages, c3.conf with dead nodes, files not synced, ) Best regards, Olivier. -- Olivier LAHAYE CEA DRT/LIST/DCSI/DIR De : Costel Seitan [csei...@slb.com] Date d'envoi : jeudi 28 février 2013 15:36 À : LAHAYE Olivier; oscar-users@lists.sourceforge.net Cc: oscar-de...@lists.sourceforge.net Objet : RE: [Oscar-users] RE : RE : RE :OSCAR on CentOS-6: how to test.(Warning, experimental) WAS:Re: OSCAR on CentOS-6: bootstrap successfull using online repository. Hello, Thank you for your comments Olivier. The nodes should not exist, however I see that the entries were added to /etc/hosts because of my previous tries: .. # OSCAR block: OSCAR hosts 192.168.0.86 epsl86 192.168.0.87 epsl87 192.168.0.88 epsl88 192.168.0.89 epsl89 But, what made me think of rsyncd was the message “rsync: failed to connect to 127.0.0.1: Connection refused (111)” There is something else that I noticed right now. For the last try I decided to name the SIS image “centos63” (compared to “centos6” just before). When analyzing the error messages it looks like some of the commands point to the old image directory (centos6) : ERROR: Impossible to detect the OS for /var/lib/systemimager/images/centos6 at /opt/sync_files/bin/sync_files line 319 main::distro_detect('dir', '/var/lib/systemimager/images/centos6') called at /opt/sync_files/bin/sync_files line 798 ERROR: Impossible to detect the distro (/var/lib/systemimager/images/centos6) at /opt/sync_files/bin/sync_files line 800. ERROR: Impossible to execute /opt/sync_files/bin/sync_files -i centos63 --force --image-only at /var/lib/oscar/packages/sis/api-post-image line 205. I will create the local epel repo and try again. There is no reason to do so but I will eventually name my SIS image “centos6” again. Kind Regards, Costel SEITAN -- Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_feb___ Oscar-users mailing list Oscar-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/oscar-users