Hi, Today after a "yum update" which did upgrade parts of my openstack installation and a reboot, I ended up with a "500 internal server error". Here's last 100 log entried in httpd/error_log centos 6.3 + epel (up to date :> ), selinux is in permissive mode.

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 01/12/2013 10:52 AM, Nux! wrote: > On 11.01.2013 21:30, Nux! wrote: >> Hi, >> >> Today after a "yum update" which did upgrade parts of my >> openstack installation and a reboot, I ended up with a "500 >> internal server error". >> >> Here's last 100 log entried in httpd/error_log >> >> centos 6.3 + epel (up to date :> ), selinux is in permissive >> mode. > > Hi, > > Apparently nova-api doesn't start anymore. The initscript reports > "OK" but a second later it's dead: openstack-nova-api dead but pid > file exists > > api.log fragment here http://fpaste.org/iZJr/ > > == 2013-01-12 11:45:20 19519 INFO nova.osapi_compute.wsgi.server > [-] (19519) wsgi starting up on http://0.0.0.0:8774/ 2013-01-12 > 11:45:20 19481 CRITICAL nova [-] [Errno 98] Address already in use > 2013-01-12 11:45:21 19488 INFO nova.service [-] Parent process has > died unexpectedly, exiting == > > But there is nothing listening on 8774 according to netstat and > lsof, so why is it complaining? > By any chance, did you an upgrade from Essex to Folsom release without doing migration steps? Another idea: Did you install OpenStack with openstack-demo-install? Cinder/nova-volume requires a Volume group named cinder-volume You might also want to stop nova-volume service and start cinder-volumes/nova-volumes. If you installed OpenStack following the TestDay page, cinder/nova-volume service is not started by default, and also a volume group is not created persistantly. Dashboard in it's current version requires a volume service enabled (and running/answering to requests!) An output of openstack-status and from keystone service-list will help us here. - -- Matthias Runge <mrunge(a)matthias-runge.de&gt; <mrunge(a)fedoraproject.org&gt; _______________________________________________ cloud mailing list cloud(a)lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/cloud

On 01/11/2013 09:30 PM, Nux! wrote: > Hi, > > Today after a "yum update" which did upgrade parts of my openstack installation and a reboot, I ended up with a "500 internal server error". What were the old and new versions of nova? > Here's last 100 log entried in httpd/error_log > > centos 6.3 + epel (up to date :> ), selinux is in permissive mode. This could be due to trying to start the metadata service twice: https://lists.launchpad.net/openstack/msg19294.html https://bugzilla.redhat.com/show_bug.cgi?id=877606 You can probably get around the issue by disabling the separate metadata service or adjusting the enabled_apis option in nova.conf

On 13.01.2013 02:11, Pádraig Brady wrote: > On 01/13/2013 01:58 AM, Pádraig Brady wrote: >> On 01/11/2013 09:30 PM, Nux! wrote: >>> Hi, >>> >>> Today after a "yum update" which did upgrade parts of my openstack >>> installation and a reboot, I ended up with a "500 internal server >>> error". >> >> What were the old and new versions of nova? >> >>> Here's last 100 log entried in httpd/error_log >>> >>> centos 6.3 + epel (up to date :> ), selinux is in permissive mode. >> >> This could be due to trying to start the metadata service twice: >> >> https://lists.launchpad.net/openstack/msg19294.html >> https://bugzilla.redhat.com/show_bug.cgi?id=877606 >> >> You can probably get around the issue by disabling >> the separate metadata service or adjusting the >> enabled_apis option in nova.conf > > By the way, did you manually enable the metadata service, > as openstack-demo-install doesn't seem to at least. > > thanks, > Pádraig. > _______________________________________________ > cloud mailing list > cloud(a)lists.fedoraproject.org > https://admin.fedoraproject.org/mailman/listinfo/cloud That was probably by mistake. :-/ The dashboard is half-functioning now (no CSS, ajax etc) and nova-compute refuses to start: 2013-01-13 10:54:32 INFO nova.compute.manager [req-9c31b571-69e4-4e07-8cb2-cdf394837b72 None None] [instance: 9aec8016-e904-41ac-80ad-ee2cf95ab393] Rebooting instance after nova-compute restart. 2013-01-13 10:54:32 INFO nova.virt.libvirt.firewall [req-9c31b571-69e4-4e07-8cb2-cdf394837b72 None None] [instance: 9aec8016-e904-41ac-80ad-ee2cf95ab393] Called setup_basic_filtering in nwfilter 2013-01-13 10:54:32 INFO nova.virt.libvirt.firewall [req-9c31b571-69e4-4e07-8cb2-cdf394837b72 None None] [instance: 9aec8016-e904-41ac-80ad-ee2cf95ab393] Ensuring static filters 2013-01-13 10:54:36 8943 CRITICAL nova [-] Requested operation is not valid: domain is already running All the VMs are of course running, I'm using "start_guests_on_host_boot=true".

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 01/13/2013 09:56 AM, Nux! wrote: > > That was probably by mistake. :-/ > > The dashboard is half-functioning now (no CSS, ajax etc) and Sounds like /static from your web root is not accessible any more. That should be a config option in /etc/httpd/conf.d/openstack-dashboard.conf Please also validate /usr/share/openstack-dashboard/static is readable. If you simply reinstall openstack-dashboard, I'd do: yum remove openstack-dashboard rm -rf /etc/openstack-dashboard rm -rf /usr/share/openstack-dashboard rm -f /etc/httpd/conf.d/openstack-dashboard.conf yum install openstack-dashboard That should solve that issue.