running "rados -p rbd lock list gateway.conf" gives:
{"objname":"gateway.conf","locks":[{"name":"lock"}]}
To be sure I stopped all related services (tcmu-runner, rbd-target-gw,
rbd-target-api) on both gateways and ran "rados -p rbd lock list
gateway.conf" again, result was the same as above. Honestly I don't know
what to do with that...
To show a more complete picture I'm posting more logs below when trying
to create another disk in gwcli with "cluster_client_name =
client.iscsi" in /etc/ceph/iscsi-gateway.cfg.
The main question is:
does the gateway work with a user other than client.admin (e.g.
client.iscsi)?
if yes: what privileges does this user need?
thx
matthias
Apr 3 15:52:31 ceiscsi0 tcmu-runner: tcmu_notify_lock_lost:222
rbd/iscsi.test01: Async lock drop. Old state 1
Apr 3 15:52:31 ceiscsi0 tcmu-runner: alua_implicit_transition:566
rbd/iscsi.test01: Starting lock acquisition operation.
Apr 3 15:52:31 ceiscsi0 tcmu-runner: tcmu_rbd_lock:762
rbd/iscsi.test01: Acquired exclusive lock.
Apr 3 15:52:44 ceiscsi0 journal: ::1 - - [03/Apr/2019 15:52:44] "GET
/api/config HTTP/1.1" 200 -
Apr 3 15:52:44 ceiscsi0 journal: ::1 - - [03/Apr/2019 15:52:44] "GET
/api/config HTTP/1.1" 200 -
Apr 3 15:52:44 ceiscsi0 journal: ::ffff:xxx.xxx.215.61 - - [03/Apr/2019
15:52:44] "GET /api/_ping HTTP/1.1" 200 -
Apr 3 15:52:44 ceiscsi0 journal: ::ffff:xxx.xxx.215.61 - - [03/Apr/2019
15:52:44] "GET /api/_ping HTTP/1.1" 200 -
Apr 3 15:52:48 ceiscsi0 tcmu-runner: tcmu_notify_lock_lost:222
rbd/iscsi.test01: Async lock drop. Old state 1
Apr 3 15:52:48 ceiscsi0 tcmu-runner: alua_implicit_transition:566
rbd/iscsi.test01: Starting lock acquisition operation.
Apr 3 15:52:49 ceiscsi0 tcmu-runner: tcmu_rbd_lock:762
rbd/iscsi.test01: Acquired exclusive lock.
Apr 3 15:52:54 ceiscsi0 journal: ::1 - - [03/Apr/2019 15:52:54] "GET
/api/_ping HTTP/1.1" 200 -
Apr 3 15:52:54 ceiscsi0 journal: (LUN.add_dev_to_lio) Adding image
'iscsi/test02' to LIO backstore user:rbd
Apr 3 15:52:54 ceiscsi0 journal: (LUN.add_dev_to_lio) Successfully
added iscsi/test02 to LIO
Apr 3 15:52:54 ceiscsi0 journal: (LUN.allocate) added 'iscsi/test02' to
LIO and config object
Apr 3 15:52:54 ceiscsi0 journal: ::ffff:xxx.xxx.215.61 - - [03/Apr/2019
15:52:54] "GET /api/_ping HTTP/1.1" 200 -
Apr 3 15:53:04 ceiscsi0 journal: ::ffff:xxx.xxx.215.61 - - [03/Apr/2019
15:53:04] "GET /api/_ping HTTP/1.1" 200 -
Apr 3 15:53:13 ceiscsi0 tcmu-runner: tcmu_notify_lock_lost:222
rbd/iscsi.test01: Async lock drop. Old state 1
Apr 3 15:53:13 ceiscsi0 tcmu-runner: alua_implicit_transition:566
rbd/iscsi.test01: Starting lock acquisition operation.
Apr 3 15:53:13 ceiscsi0 tcmu-runner: tcmu_rbd_lock:762
rbd/iscsi.test01: Acquired exclusive lock.
Apr 3 15:53:14 ceiscsi0 journal: ::ffff:xxx.xxx.215.61 - - [03/Apr/2019
15:53:14] "GET /api/_ping HTTP/1.1" 200 -
Apr 3 15:53:24 ceiscsi0 journal: ::ffff:xxx.xxx.215.61 - - [03/Apr/2019
15:53:24] "GET /api/_ping HTTP/1.1" 200 -
Apr 3 15:53:24 ceiscsi0 journal: (Config.lock) Timed out (30s) waiting
for excl lock on gateway.conf object
Apr 3 15:53:24 ceiscsi0 journal: LUN alloc problem - Timed out (30s)
waiting for excl lock on gateway.conf object
Apr 3 15:53:24 ceiscsi0 journal: ::1 - - [03/Apr/2019 15:53:24] "PUT
/api/_disk/iscsi/test02 HTTP/1.1" 500 -
Apr 3 15:53:24 ceiscsi0 journal: _disk change on localhost failed with 500
Apr 3 15:53:24 ceiscsi0 journal: ::1 - - [03/Apr/2019 15:53:24] "PUT
/api/disk/iscsi/test02 HTTP/1.1" 500 -
Apr 3 15:53:34 ceiscsi0 tcmu-runner: tcmu_notify_lock_lost:222
rbd/iscsi.test01: Async lock drop. Old state 1
Apr 3 15:53:34 ceiscsi0 tcmu-runner: alua_implicit_transition:566
rbd/iscsi.test01: Starting lock acquisition operation.
Apr 3 15:53:34 ceiscsi0 journal: ::ffff:xxx.xxx.215.61 - - [03/Apr/2019
15:53:34] "GET /api/_ping HTTP/1.1" 200 -
Apr 3 15:53:35 ceiscsi0 tcmu-runner: tcmu_rbd_lock:762
rbd/iscsi.test01: Acquired exclusive lock.
Am 01.04.19 um 17:05 schrieb Jason Dillaman:
What happens when you run "rados -p rbd lock list gateway.conf"?
On Fri, Mar 29, 2019 at 12:19 PM Matthias Leopold
<matthias.leop...@meduniwien.ac.at> wrote:
Hi,
I upgraded my test Ceph iSCSI gateways to
ceph-iscsi-3.0-6.g433bbaa.el7.noarch.
I'm trying to use the new parameter "cluster_client_name", which - to me
- sounds like I don't have to access the ceph cluster as "client.admin"
anymore. I created a "client.iscsi" user and watched what happened. The
gateways can obviously read the config (which I created when I was still
client.admin), but when I try to change anything (like create a new disk
in pool "iscsi") I get the following error:
(Config.lock) Timed out (30s) waiting for excl lock on gateway.conf object
I suspect this is related to the privileges of "client.iscsi", but I
couldn't find the correct settings yet. The last thing I tried was:
caps: [mon] allow r, allow command "osd blacklist"
caps: [osd] allow * pool=rbd, profile rbd pool=iscsi
Can anybody tell me how to solve this?
My Ceph version is 12.2.10 on CentOS 7.
thx
Matthias
_______________________________________________
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
--
Matthias Leopold
IT Systems & Communications
Medizinische Universität Wien
Spitalgasse 23 / BT 88 /Ebene 00
A-1090 Wien
Tel: +43 1 40160-21241
Fax: +43 1 40160-921200
_______________________________________________
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com