Hi Josy, this is correct.
Just make sure that your current user as well as the user for your VMs (if you are using a VM environment) are allowed to write to this directory. Also make sure that /var/run/ceph exists. Once you have fixed the permissions problem and made sure that the path where you want to create the socket file exists it will be OK. Note that the socket file can be created anywhere. So you could actually position the parameter as admin_socket = “./my-client.asok” just for the purpose of a test Regards JC Lopez Senior Technical Instructor, Global Storage Consulting Practice Red Hat, Inc. jelo...@redhat.com <mailto:jelo...@redhat.com> +1 408-680-6959 > On Oct 17, 2017, at 16:07, Josy <j...@colossuscloudtech.com> wrote: > > I think it is permission error, because when running ceph -s it shows this > error at the top > > ------------- > > $ ceph -s > 2017-10-17 15:53:26.132180 7f7698834700 -1 asok(0x7f76940017a0) > AdminSocketConfigObs::init: failed: AdminSocket::bind_and_listen: failed to > bind the UNIX domain socket to > '/var/run/ceph/ceph-client.admin.29983.140147265902928.asok': (13) Permission > denied > cluster: > id: de296604-d85c-46ab-a3af-add3367f0e6d > health: HEALTH_OK > ------------ > > Selinux is disabled in the server. Also I changed ownership of /var/run/ceph > to the ceph user. > Still no luck. 'ls /var/run/ceph/ lists' no files in the client server > > > > On 18-10-2017 04:07, Jason Dillaman wrote: >> On Tue, Oct 17, 2017 at 6:30 PM, Josy <j...@colossuscloudtech.com> wrote: >>> Hi, >>> >>> I am running the command from the admin server. >>> >>> Because there are no asok file in the client server >>> ls /var/run/ceph/ lists no files in the client server. >> Most likely a permissions or SElinux/AppArmor issue where the librbd >> client application cannot write to the directory. >> >>>>> As Jason points it out you also need to make sure that your restart the >>>>> client connection for the changes in the ceph.conf file to take effect. >>> You mean restart the client server ? >>> >>> (I am sorry, this is something new for me. I have just started learning >>> ceph.) >> Assuming this is for QEMU, QEMU is the librbd client so you would have >> to stop/start the VM to pick up any configuration changes (or perform >> a live migration to another server). >> >>> On 18-10-2017 03:32, Jean-Charles Lopez wrote: >>> >>> Hi Josy, >>> >>> just a doubt but it looks like your ASOK file is the one from a Ceph >>> Manager. So my suspicion is that you may be running the command from the >>> wrong machine. >>> >>> To run this command, you need to ssh into the machine where the client >>> connection is being initiated. >>> >>> But may be I am wrong regarding your exact connection point. >>> >>> As Jason points it out you also need to make sure that your restart the >>> client connection for the changes in the ceph.conf file to take effect. >>> >>> Regards >>> JC Lopez >>> Senior Technical Instructor, Global Storage Consulting Practice >>> Red Hat, Inc. >>> jelo...@redhat.com >>> +1 408-680-6959 >>> >>> On Oct 17, 2017, at 14:29, Josy <j...@colossuscloudtech.com> wrote: >>> >>> Thanks for the reply. >>> >>> I added rbd_non_blocking_aio = false in ceph.conf and pushed the admin file >>> to all nodes. >>> >>> ----- >>> [client] >>> admin socket = /var/run/ceph/$cluster-$type.$id.$pid.$cctid.asok >>> log file = /var/log/ceph/client.log >>> debug rbd = 20 >>> debug librbd = 20 >>> rbd_non_blocking_aio = false >>> ------ >>> >>> >>> However the config show command still shows it as true. >>> ----------- >>> [cephuser@ceph-las-admin-a1 ceph-cluster]$ sudo ceph --admin-daemon >>> /var/run/ceph/ceph-mgr.ceph-las-admin-a1.asok config show | grep >>> "rbd_non_blocking_aio" >>> "rbd_non_blocking_aio": "true", >>> ----------- >>> >>> Did I miss something ? >>> >>> >>> On 18-10-2017 01:22, Jean-Charles Lopez wrote: >>> >>> Hi >>> >>> syntax uses the admin socket file : ceph --admin-daemon >>> /var/run/ceph/$cluster-$type.$id.$pid.$cctid.asok config get rbd_cache >>> >>> Should be /var/run/ceph/ceph.client.admin.$pid.$cctid.asok if your >>> connection is using client.admin to connect to the cluster and your cluster >>> name is set to the default of ceph. But obviously can’t know from here the >>> PID and the CCTID you will have to identify. >>> >>> You can actually do a ls /var/run/ceph to find the correct admin socket file >>> >>> Regards >>> JC Lopez >>> Senior Technical Instructor, Global Storage Consulting Practice >>> Red Hat, Inc. >>> jelo...@redhat.com >>> +1 408-680-6959 >>> >>> On Oct 17, 2017, at 12:50, Josy <j...@colossuscloudtech.com> wrote: >>> >>> Hi, >>> >>> >>> I am following this article : >>> >>> http://ceph.com/geen-categorie/ceph-validate-that-the-rbd-cache-is-active/ >>> >>> I have enabled this flag in ceph.conf >>> >>> [client] >>> admin socket = /var/run/ceph/$cluster-$type.$id.$pid.$cctid.asok >>> log file = /var/log/ceph/ >>> >>> >>> But the command to show the conf is not working : >>> >>> [cephuser@ceph-las-admin-a1 ceph-cluster]$ sudo ceph --admin-daemon >>> /etc/ceph/ceph.client.admin.keyring config show >>> admin_socket: exception getting command descriptions: [Errno 111] Connection >>> refused >>> >>> _______________________________________________ >>> ceph-users mailing list >>> ceph-users@lists.ceph.com >>> http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com >>> >>> >>> >>> >>> >>> >>> _______________________________________________ >>> ceph-users mailing list >>> ceph-users@lists.ceph.com >>> http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com >>> >> >> > >
_______________________________________________ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
