Hi Adam,
Big Thanks for the responses and clarifying the global usage of the --image
parameter. Eventhough, I gave --image during bootstrap only mgr & mon
daemons on the bootstrap host are getting created with that image and the
rest of the demons are created on the image daemon-base as I mentioned
earlier.
So, there are two images coming into action here. First one can be
controlled with the --image parameter in bootstrap( which worked when
supplied in front of bootstrap keyword).
The second container image is controlled by the variable 'container_image'
which is set to 'docker.io/ceph/daemon-base:latest-pacific-devel' by
default.
Even Though it can be modified at runtime after bootstrap the existing
daemons will not be modified. But, that case can be handled with the 'ceph
orch upgrade' command like you mentioned at first.
However, it is observed that if we mention this variable in the bootstrap
config file, all the daemons will be created with the mentioned image from
bootstrap itself.
So, the take away from this is we mention the first image using the
'--image' argument in bootstrap command and second image using the variable
'container_image' in the bootstrap config file all dameons will be created
with the same image.
So the question is, is cephadm really require two images?
Also, one more observation I had is; even though I gave the same image in
both the above said provisions, I can see a difference among the same type
of daemons created on different hosts. (Even Though all dameons uses single
image in effect)
Following is a result of a cluster created on 3 hosts. The bootstrap
command is below:(rest of the services are deployed using ceph orch)
'sudo cephadm --image quay.io/ceph/ceph:v16.2.7 bootstrap
--skip-monitoring-stack --mon-ip 10.175.41.11 --cluster-network
10.175.42.0/24 --ssh-user ceph_deploy --ssh-private-key
/home/ceph_deploy/.ssh/id_rsa --ssh-public-key
/home/ceph_deploy/.ssh/id_rsa.pub --config
/home/ceph_deploy/ceph_bootstrap/ceph.conf --initial-dashboard-password
Qwe4Rt6D33 --dashboard-password--noupdate --no-minimize-config
[root@hcictrl01 stack_orchestrator]# ceph orch ls
NAME PORTS RUNNING REFRESHED AGE PLACEMENT
crash 3/3 9m ago 15m *
mds.cephfs 3/3 9m ago 9m
hcictrl02;hcictrl03;hcictrl01;count:3
mgr 3/3 9m ago 13m
hcictrl02;hcictrl03;hcictrl01;count:3
mon 3/5 9m ago 15m count:5
osd 8 9m ago - <unmanaged>
rgw.rgw ?:7480 3/3 9m ago 9m
hcictrl02;hcictrl03;hcictrl01;count:3
[root@hcictrl01 stack_orchestrator]# ceph orch ps
NAME HOST PORTS STATUS REFRESHED
AGE MEM USE MEM LIM VERSION IMAGE ID CONTAINER ID
crash.hcictrl01 hcictrl01 running (15m) 9m ago
15m 6983k - 16.2.7 231fd40524c4 f6f866f4be92
crash.hcictrl02 hcictrl02 running (14m) 9m ago
14m 6987k - 16.2.7 231fd40524c4 1cb62e191c07
crash.hcictrl03 hcictrl03 running (14m) 9m ago
14m 6995k - 16.2.7 231fd40524c4 3e03f99065c0
mds.cephfs.hcictrl01.vuamjy hcictrl01 running (10m) 9m ago
10m 13.0M - 16.2.7 231fd40524c4 9b3aeab68115
mds.cephfs.hcictrl02.myohpi hcictrl02 running (10m) 9m ago
10m 15.6M - 16.2.7 231fd40524c4 5cded1208028
mds.cephfs.hcictrl03.jziler hcictrl03 running (10m) 9m ago
10m 12.6M - 16.2.7 231fd40524c4 94dccd01a123
mgr.hcictrl01.ljtznv hcictrl01 *:9283 running (16m) 9m ago
16m 428M - 16.2.7 231fd40524c4 5bc89cc72b37
mgr.hcictrl02.izfnvh hcictrl02 *:8443 running (14m) 9m ago
14m 382M - 16.2.7 231fd40524c4 acd435a8b6b1
mgr.hcictrl03.eekrgo hcictrl03 *:8443 running (14m) 9m ago
14m 382M - 16.2.7 231fd40524c4 3c241b35a3fe
mon.hcictrl01 hcictrl01 running (16m) 9m ago
16m 79.9M 2048M 16.2.7 231fd40524c4 19a2db98043a
mon.hcictrl02 hcictrl02 running (14m) 9m ago
14m 88.2M 2048M 16.2.7 231fd40524c4 4941ee015e75
mon.hcictrl03 hcictrl03 running (14m) 9m ago
14m 86.2M 2048M 16.2.7 231fd40524c4 150cb8e9d25d
osd.0 hcictrl02 running (13m) 9m ago
13m 31.8M 1536M 16.2.7 231fd40524c4 9dc98cc6ba3d
osd.1 hcictrl02 running (13m) 9m ago
13m 42.4M 1536M 16.2.7 231fd40524c4 62caab356b02
osd.2 hcictrl02 running (12m) 9m ago
12m 58.9M 1536M 16.2.7 231fd40524c4 38ea17598fa5
osd.3 hcictrl01 running (12m) 9m ago
12m 35.1M 1536M 16.2.7 231fd40524c4 03284b111258
osd.4 hcictrl01 running (11m) 9m ago
11m 44.1M 1536M 16.2.7 231fd40524c4 3e3315fbb46a
osd.5 hcictrl01 running (11m) 9m ago
11m 54.6M 1536M 16.2.7 231fd40524c4 1f90169412a5
osd.6 hcictrl03 running (11m) 9m ago
11m 35.2M 1536M 16.2.7 231fd40524c4 d26811c0c9fc
osd.7 hcictrl03 running (10m) 9m ago
10m 59.1M 1536M 16.2.7 231fd40524c4 1ce10dcfa6ca
rgw.rgw.hcictrl01.ceruvj hcictrl01 *:7480 running (9m) 9m ago
9m 49.6M - 16.2.7 231fd40524c4 c2c45767cda5
rgw.rgw.hcictrl02.pzgwht hcictrl02 *:7480 running (9m) 9m ago
9m 18.6M - 16.2.7 231fd40524c4 461c0869e559
rgw.rgw.hcictrl03.ryrptr hcictrl03 *:7480 running (9m) 9m ago
9m 20.9M - 16.2.7 231fd40524c4 8fac107de350
[root@hcictrl01 stack_orchestrator]# ceph config dump | grep -i image
global basic container_image
quay.io/ceph/ceph@sha256:ac9f32e9a27ded104c2ed64ddbe112d67a461edafb6a7a53525e72fc69df759e
*
container list on bootstrap host:(image name is not uniform across all
dameons)
[root@hcictrl01 stack_orchestrator]# podman ps
CONTAINER ID IMAGE
COMMAND CREATED STATUS
PORTS NAMES
19a2db98043a quay.io/ceph/ceph:v16.2.7
-n mon.hcictrl01 ... 17 minutes ago Up 17
minutes ago
ceph-c5aa753a-8422-11ec-b231-0015171590ba-mon-hcictrl01
5bc89cc72b37 quay.io/ceph/ceph:v16.2.7
-n mgr.hcictrl01.... 17 minutes ago Up 17
minutes ago
ceph-c5aa753a-8422-11ec-b231-0015171590ba-mgr-hcictrl01-ljtznv
f6f866f4be92
quay.io/ceph/ceph@sha256:ac9f32e9a27ded104c2ed64ddbe112d67a461edafb6a7a53525e72fc69df759e
-n client.crash.h... 15 minutes ago Up 15 minutes ago
ceph-c5aa753a-8422-11ec-b231-0015171590ba-crash-hcictrl01
03284b111258
quay.io/ceph/ceph@sha256:ac9f32e9a27ded104c2ed64ddbe112d67a461edafb6a7a53525e72fc69df759e
-n osd.3 -f --set... 12 minutes ago Up 12 minutes ago
ceph-c5aa753a-8422-11ec-b231-0015171590ba-osd-3
3e3315fbb46a
quay.io/ceph/ceph@sha256:ac9f32e9a27ded104c2ed64ddbe112d67a461edafb6a7a53525e72fc69df759e
-n osd.4 -f --set... 12 minutes ago Up 12 minutes ago
ceph-c5aa753a-8422-11ec-b231-0015171590ba-osd-4
1f90169412a5
quay.io/ceph/ceph@sha256:ac9f32e9a27ded104c2ed64ddbe112d67a461edafb6a7a53525e72fc69df759e
-n osd.5 -f --set... 11 minutes ago Up 11 minutes ago
ceph-c5aa753a-8422-11ec-b231-0015171590ba-osd-5
9b3aeab68115
quay.io/ceph/ceph@sha256:ac9f32e9a27ded104c2ed64ddbe112d67a461edafb6a7a53525e72fc69df759e
-n mds.cephfs.hci... 10 minutes ago Up 10 minutes ago
ceph-c5aa753a-8422-11ec-b231-0015171590ba-mds-cephfs-hcictrl01-vuamjy
c2c45767cda5
quay.io/ceph/ceph@sha256:ac9f32e9a27ded104c2ed64ddbe112d67a461edafb6a7a53525e72fc69df759e
-n client.rgw.rgw... 10 minutes ago Up 10 minutes ago
ceph-c5aa753a-8422-11ec-b231-0015171590ba-rgw-rgw-hcictrl01-ceruvj
In the above output, the mon and mgr on bootstrap node is referring to the
container image with tag (v16.2.7 ) unlike the rest of the daemons where it
refers with the image digest.
Even Though the digest belong to the same image, a non-uniform behaviour is
observed in the warning generated by mons(see below)
[root@hcictrl01 stack_orchestrator]# ceph health detail
HEALTH_WARN mons are allowing insecure global_id reclaim
[WRN] AUTH_INSECURE_GLOBAL_ID_RECLAIM_ALLOWED: mons are allowing insecure
global_id reclaim
mon.hcictrl02 has auth_allow_insecure_global_id_reclaim set to true
mon.hcictrl03 has auth_allow_insecure_global_id_reclaim set to true
