First time here? Check out the FAQ!
2018-11-17 13:52:42 -0600 | received badge | ● Famous Question (source) |
2018-06-27 21:25:48 -0600 | received badge | ● Famous Question (source) |
2018-06-27 21:25:48 -0600 | received badge | ● Famous Question (source) |
2018-06-09 05:01:04 -0600 | received badge | ● Famous Question (source) |
2018-05-24 21:20:07 -0600 | answered a question | How to configure huge pages of 1G manually on computes? grub2-mkconfig -o /etc/grub2.cfg You have to rebuild the /boot/grub file which the above command will do. |
2018-05-22 16:54:10 -0600 | received badge | ● Famous Question (source) |
2018-05-10 00:54:54 -0600 | received badge | ● Notable Question (source) |
2018-05-07 02:52:21 -0600 | received badge | ● Popular Question (source) |
2018-05-03 11:05:38 -0600 | asked a question | controller.pp fails with "Error: Failed to apply catalog: Execution of '/usr/bin/openstack domain list --quiet --format csv' returned 1: Service Unavailable (HTTP 503)" I installed this on a VM at home and it installs just fine. No idea why this would fail on a 32 core server with 64G of ram. This system is behind a proxy. |
2018-05-03 11:02:17 -0600 | answered a question | In TripleO install, the compute nodes default route is always to the undercloud IP? Why does the host not get a direct route to the external network for things like NTP, external Ceph, and DNS? Old question but the answer is to use the Management network. That is what it is for. |
2018-05-03 11:01:40 -0600 | received badge | ● Notable Question (source) |
2017-10-17 06:55:48 -0600 | received badge | ● Popular Question (source) |
2017-07-16 04:01:22 -0600 | received badge | ● Famous Question (source) |
2017-07-15 08:50:57 -0600 | received badge | ● Notable Question (source) |
2017-07-14 07:11:13 -0600 | received badge | ● Popular Question (source) |
2017-07-12 22:27:07 -0600 | received badge | ● Notable Question (source) |
2017-06-20 15:04:58 -0600 | answered a question | Enable the display of the Metadata definitions catalog in Glance Found it: "glance-manage db_load_metadefs" That loads the system defaults. |
2017-06-20 10:43:36 -0600 | asked a question | Enable the display of the Metadata definitions catalog in Glance I saw this youtube video showing this https://www.youtube.com/watch?v=zJpHXdBOoeM (OpenStack Metadata Definitions Concept Demo) I can't tell if this is meant to available and if so how? In RHOSP10 Newton, I go the the tab mentioned for Metadata and I don't see any definitions. I'd like to enable them as we are moving to OVS-DPDK and these are going to be very useful in defining new flavors and launching images. |
2017-06-15 00:19:44 -0600 | received badge | ● Popular Question (source) |
2017-06-15 00:19:16 -0600 | answered a question | Can anyone explain the templates used in the NFV architecture of RHOSP 10 or RHOSP11? FYI, I am well on my way to installing RHOSP 10 with DPDK and no, they are not correct. I have a bug opened with RHEL to get it corrected. RH has midified the current templates and the issue is resolved. |
2017-06-15 00:17:52 -0600 | received badge | ● Supporter (source) |
2017-06-15 00:16:43 -0600 | commented answer | how instance storage work? Is the above description still correct with Volume Boot from Ceph? |
2017-06-13 19:43:24 -0600 | received badge | ● Popular Question (source) |
2017-06-13 14:50:44 -0600 | asked a question | Can anyone explain the templates used in the NFV architecture of RHOSP 10 or RHOSP11? The drawings and templates don't appear to match. For instance, there is no configuration of NIC1 with the IP of the undercloud. Also, the external network is not assigned in the controller. Both seem odd to me. |
2017-06-13 14:47:03 -0600 | answered a question | Launching instance on RHOSP 10 (Newton) fails with memory over 30G Found to be a bug in RHOSP 10 that has been fixed. Apply the latest patch |
2017-06-08 17:30:29 -0600 | received badge | ● Enthusiast |
2017-06-05 17:29:21 -0600 | asked a question | Launching instance on RHOSP 10 (Newton) fails with memory over 30G When launching an instance I am getting the below failure: Has anyone had experience with this. This is a DPDK system with multiple kernel command line options used. |
2017-05-24 14:30:16 -0600 | commented answer | In Newton and beyond, do we need to set the schedular filters specifically or is there a generic set of defaults? So you are saying I only need to call out the additional filters I need? |
2017-05-24 14:26:17 -0600 | answered a question | Problem during installation Newton on Centos7 That sounds like an issue with your profiles. Is this Mitaka or Newton? If Newton, try openstack profile list and make sure they are in the correct order. |
2017-05-22 09:20:20 -0600 | asked a question | In TripleO install, the compute nodes default route is always to the undercloud IP? Why does the host not get a direct route to the external network for things like NTP, external Ceph, and DNS? Question stated in Title |
2017-05-22 09:17:02 -0600 | asked a question | In Newton and beyond, do we need to set the schedular filters specifically or is there a generic set of defaults? In Nova.conf on the controller, I am specifically setting the default filters as follows: However, on all my other Newton installs, I have not needed "NumaTopology". So, I don't set any config and things like Aggregate and Affinity Filters are there. In the docs, it seems to always say the scheduler default filters config line needs to always be there. What is the correct way to do this? |
2017-05-11 18:29:44 -0600 | received badge | ● Popular Question (source) |
2017-05-11 18:29:44 -0600 | received badge | ● Notable Question (source) |
2017-01-23 06:56:15 -0600 | asked a question | Can't launch instance after adding disk space to /var/log/nova/instances I had an issue when I was launching instances due to free space in /var/log/nova/instances and I had a free drive available on the compute nodes. So, I formatted and mounted /dev/sdb1 under /var/log/nova/instances. This fixed the error I was having starting large instances ("no node available") but now all launches fail when opening "console.log". What I found is that all files under /var/log/nova/instances are not owned by user id nova, group id nova but are also owned by qemu. The qemu files are the ones failing (console.log). I have checked other openstack installs I have done and the permissions appear exactly the same. Any help setting permissions correctly would be appreciated. I could reinstall but I do not see the need to do so for something as simple as this. "/var/lib/nova/instances/d9ca9e6e-57cd-4127-b1e4-0edc4fa46160/console.log: Permission denied" Taken from a working system below: Not Working compute node: (more) |