Bug #55160
closedFailed to deploy iscsi gateway with orch/dashboard
0%
Description
Hi, my ceph version is 16.2.7 and I'm following the doc from: https://docs.ceph.com/en/latest/cephadm/services/iscsi to deploy iscsi service.
However, I'm failed when I add iscsi target. the current state is:
[root@ceph-1 /]# gwcli
1 gateway is inaccessible - updates will be disabled
/* omitted */
o- iscsi-targets ............................................................................... [DiscoveryAuth: None, Targets: 1]
o- iqn.2001-07.com.ceph:1648746014395 ................................................................ [Auth: None, Gateways: 1]
o- disks .......................................................................................................... [Disks: 0]
o- gateways ............................................................................................ [Up: 0/1, Portals: 1]
| o- ceph-1 ....................................................................................... [10.18.1.1 (UNAUTHORIZED)]
o- host-groups .................................................................................................. [Groups : 0]
o- hosts ....................................................................................... [Auth: ACL_ENABLED, Hosts: 0]
it shows that my gateway is UNAUTHORIZED but I don't know why.
I also can't delete the added gateway even with force flag:
/iscsi-target...4395/gateways> delete ceph-1
Deleting gateway, ceph-1
Deleting the last gateway will remove all objects on this target. Use confirm=true
/iscsi-target...4395/gateways> delete ceph-1 confirm=true
Deleting gateway, ceph-1
Could not contact ceph-1. If the gateway is permanently down. Use confirm=true to force removal. WARNING: Forcing removal of a gateway that can still be reached by an initiator may result in data corruption.
/iscsi-target...4395/gateways>
So I'm stucking here.
And I've done more investigation:
from the rbd-target-api, I got the log from above delete process:
debug fe80::1314:bf0d:d4be:717b%admin - - [01/Apr/2022 15:11:11] "GET /api/_ping HTTP/1.1" 403 -
debug ::1 - - [01/Apr/2022 15:11:11] "DELETE /api/gateway/iqn.2001-07.com.ceph:1648746014395/ceph-1 HTTP/1.1" 503 -
Only 503 status code with on other information. since 503 usually stands for service unavailable. I've notice that there might be another `service` which is not running properly.
and after read the manual deployment doc from: https://docs.ceph.com/en/latest/rbd/iscsi-target-cli/ I suddenly realized that it doesn't start rbd-target-gw process on my ceph cluster.
Did I miss something and how I can get these stuff work?
Thanks
Updated by Jiayu Sun about 2 years ago
after read some code from: https://github.com/ceph/ceph-iscsi
I found that gateway names (like ceph-1) should be resolved to the correct ip address.
I'm going to add entries in /etc/hosts. but which ip should i use for hostname? because I have 2 ip address for each node of cluster(e.g. cluster network/public network)
Updated by Xiubo Li about 2 years ago
Jiayu Sun wrote:
after read some code from: https://github.com/ceph/ceph-iscsi
I found that gateway names (like ceph-1) should be resolved to the correct ip address.
I'm going to add entries in /etc/hosts. but which ip should i use for hostname? because I have 2 ip address for each node of cluster(e.g. cluster network/public network)
This should be the public IPs.
Updated by Jiayu Sun about 2 years ago
Xiubo Li wrote:
Jiayu Sun wrote:
after read some code from: https://github.com/ceph/ceph-iscsi
I found that gateway names (like ceph-1) should be resolved to the correct ip address.
I'm going to add entries in /etc/hosts. but which ip should i use for hostname? because I have 2 ip address for each node of cluster(e.g. cluster network/public network)
This should be the public IPs.
Thanks Xiubo, It works now!
All I have to do is make sure the hostname of each cluster can be resovled to public ip address on each node. We can close this tick
Updated by Ilya Dryomov about 2 years ago
- Status changed from In Progress to Closed