now you can use a network link name with a hyphen #1045

wind0204 · 2016-10-24T15:53:47Z

leseb · 2016-10-25T08:46:49Z

Have you been able to test and validate that?
It'd be nice if you could share some logs :), thanks!

wind0204 · 2016-10-25T11:03:42Z

@leseb sure; though I've gotten stuck at another issue ( #760 )

see my a little weak proof:

root@0-h0:~/ceph-ansible# cat /etc/ceph/ceph.conf
[global]
mon host = 192.168.255.227,192.168.255.254
...

root@peta-s-227:~/mnt# cat /etc/ceph/ceph.conf
[global]
mon host = 192.168.255.227,192.168.255.254
...

other settings are the same with the issue thread ( #1038 )

leseb · 2016-10-25T12:54:16Z

How many mons do you have? Only 2?
If you can't collect keys, this means monitors didn't start or are not in quorum.
Please check if processes are running, if so check their logs.

wind0204 · 2016-10-25T14:24:09Z

@leseb I have 2 monitors and they seem to be up and running:

root@peta-s-227:~/mnt# systemctl status ceph-mon@*
● [email protected] - Ceph cluster monitor daemon
   Loaded: loaded (/lib/systemd/system/[email protected]; enabled)
   Active: active (running) since Tue 2016-10-25 20:02:55 KST; 3h 17min ago
 Main PID: 25182 (ceph-mon)
   CGroup: /system.slice/system-ceph\x2dmon.slice/[email protected]
           └─25182 /usr/bin/ceph-mon -f --cluster ceph --id peta-s-227 --setuser ceph --setgroup ceph

Oct 25 20:02:55 peta-s-227 systemd[1]: Started Ceph cluster monitor daemon.
Oct 25 20:02:57 peta-s-227 ceph-mon[25182]: starting mon.peta-s-227 rank -1 at 192.168.255.227:6789/0 mon_data /var/lib/ceph/mon/ceph-peta-s-227 f...5b39fc6bf
Oct 25 23:20:28 peta-s-227 systemd[1]: [/lib/systemd/system/[email protected]:24] Unknown lvalue 'TasksMax' in section 'Service'
Hint: Some lines were ellipsized, use -l to show in full.
root@peta-s-227:~/mnt# ps -el | grep ceph-mon
4 S 64045 25182     1 30  80   0 - 3776679 -    ?        01:00:08 ceph-mon

doeals0@0-h0:~$ systemctl status ceph-mon@*
● [email protected] - Ceph cluster monitor daemon
   Loaded: loaded (/lib/systemd/system/[email protected]; enabled)
   Active: active (running) since Tue 2016-10-25 20:53:04 KST; 2h 29min ago
 Main PID: 25945 (ceph-mon)
   CGroup: /system.slice/system-ceph\x2dmon.slice/[email protected]
           └─25945 /usr/bin/ceph-mon -f --cluster ceph --id 0-h0 --setuser ceph --setgroup ceph
doeals0@0-h0:~$ ps -el | grep ceph-mon
4 S 64045 25945     1 19  80   0 - 2538652 -    ?        00:29:18 ceph-mon

leseb · 2016-10-25T14:31:32Z

2 doesn't make a quorum. Drop down to 1 or go with 3 at least.

wind0204 · 2016-10-25T14:42:52Z

@leseb I could run a ceph cluster with only 2 monitors when I used ceph-deploy; what makes you say so? I don't have enough resources to deploy 3 monitors; nor do I want to have only 1 monitor risking the availability.

leseb · 2016-10-25T18:24:33Z

No no no. Monitors have quorum that can only be resolved by an odd number
of machines. Having 1 or 2 mons is the same. If you loose one of the two
the remaining one will go stale. So please DON'T DO THIS.

On Tuesday, 25 October 2016, Nicholas Gim [email protected] wrote:

@leseb https://github.com/leseb I could run a ceph cluster with only 2
monitors when I used ceph-deploy; what makes you say so? I don't have
enough resources to deploy 3 monitors; nor do I want to have only 1 monitor
risking the availability.

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#1045 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AA3tX-c-GBgj1ll8lBxo47EAdS8P9f0eks5q3hVugaJpZM4Ke9ol
.

Cheers

––––––
Sébastien Han
Senior Cloud Architect

"Always give 100%. Unless you're giving blood."

Mail: [email protected]
Address: 11 bis, rue Roquépine - 75008 Paris

wind0204 · 2016-10-26T05:43:34Z

@leseb alright, thanks a lot for the information. I've reduced it to 1 from 2; the monitor is up and running ( http://docs.ceph.com/docs/jewel/rados/configuration/mon-config-ref/#initial-members )

now I get another error message for ceph-osds this time ( related pull request : #1035 (comment) )

TASK [ceph-osd : include] ******************************************************
fatal: [peta-s-225]: FAILED! => {"failed": true, "reason": "no action detected in task. This often indicates a misspelled module name, or incorrect module path.\n\nThe error appears to have been in '/home/doeals0/ceph-ansible/roles/ceph-osd/tasks/pre_requisite.yml': line 8, column 3, but may\nbe elsewhere in the file depending on the exact syntax problem.\n\nThe offending line appears to be:\n\n\n- name: enable extras repo on centos\n  ^ here\n\n\nThe error appears to have been in '/home/doeals0/ceph-ansible/roles/ceph-osd/tasks/pre_requisite.yml': line 8, column 3, but may\nbe elsewhere in the file depending on the exact syntax problem.\n\nThe offending line appears to be:\n\n\n- name: enable extras repo on centos\n  ^ here\n"}
fatal: [peta-s-226]: FAILED! => {"failed": true, "reason": "no action detected in task. This often indicates a misspelled module name, or incorrect module path.\n\nThe error appears to have been in '/home/doeals0/ceph-ansible/roles/ceph-osd/tasks/pre_requisite.yml': line 8, column 3, but may\nbe elsewhere in the file depending on the exact syntax problem.\n\nThe offending line appears to be:\n\n\n- name: enable extras repo on centos\n  ^ here\n\n\nThe error appears to have been in '/home/doeals0/ceph-ansible/roles/ceph-osd/tasks/pre_requisite.yml': line 8, column 3, but may\nbe elsewhere in the file depending on the exact syntax problem.\n\nThe offending line appears to be:\n\n\n- name: enable extras repo on centos\n  ^ here\n"}

NO MORE HOSTS LEFT *************************************************************
        to retry, use: --limit @/home/doeals0/ceph-ansible/site.retry

PLAY RECAP *********************************************************************
0-h0                       : ok=56   changed=0    unreachable=0    failed=0   
peta-s-225                 : ok=45   changed=0    unreachable=0    failed=1   
peta-s-226                 : ok=44   changed=0    unreachable=0    failed=1

wind0204 · 2016-10-26T06:57:09Z

@leseb I got through the running of 'ansible-playbook site.yml' by negating the pull request #1035 (comment) and I presume that my pull request would help work; though I'm entirely new to Jinja2 and the other stuff so not very sure about the safety of my code.

PLAY RECAP *********************************************************************
0-h0                       : ok=56   changed=0    unreachable=0    failed=0   
peta-s-225                 : ok=65   changed=0    unreachable=0    failed=0   
peta-s-226                 : ok=64   changed=0    unreachable=0    failed=0   

doeals0@0-h0:~/ceph-ansible$ ssh ansibled@0-h0 "sudo ceph osd tree"                                                                                           
ID WEIGHT  TYPE NAME           UP/DOWN REWEIGHT PRIMARY-AFFINITY 
-1 4.09140 root default                                          
-2 2.04570     host peta-s-226                                   
 0 0.68190         osd.0            up  1.00000          1.00000 
 2 0.68190         osd.2            up  1.00000          1.00000 
 5 0.68190         osd.5            up  1.00000          1.00000 
-3 2.04570     host peta-s-225                                   
 1 0.68190         osd.1            up  1.00000          1.00000 
 3 0.68190         osd.3            up  1.00000          1.00000 
 4 0.68190         osd.4            up  1.00000          1.00000 

doeals0@0-h0:~/ceph-ansible$ ssh ansibled@0-h0 "sudo ceph status"                                                                                             
    cluster afe5c536-0d76-4dcc-89ed-b1c5b39fc6bf
     health HEALTH_WARN
            8 pgs degraded
            64 pgs stuck unclean
            8 pgs undersized
            too few PGs per OSD (21 < min 30)
     monmap e1: 1 mons at {0-h0=192.168.255.254:6789/0}
            election epoch 5, quorum 0 0-h0
     osdmap e25: 6 osds: 6 up, 6 in; 56 remapped pgs
            flags sortbitwise
      pgmap v56: 64 pgs, 1 pools, 0 bytes data, 0 objects
            201 MB used, 4189 GB / 4189 GB avail
                  31 active+remapped
                  25 active
                   8 active+undersized+degraded

leseb · 2016-10-31T10:17:19Z

Alright thanks for the comments, the code looks good to me.
Let me give it a try myself and I'll let you know.

Thanks!

leseb · 2016-10-31T15:52:37Z

test this please

leseb · 2016-10-31T15:53:33Z

LGTM, @font do you mind testing this against a containerized deployment?
Thanks!

font · 2016-10-31T18:18:47Z

@leseb It should work for containerized deployments as this change only applies when not mon_containerized_deployment and not mon_containerized_deployment_with_kv. I will test anyway to make sure.

font · 2016-10-31T18:33:29Z

@leseb Tests on containerized deployment looks good.

leseb · 2016-10-31T20:15:54Z

@font thanks a lot!

now you can use a network link name with a hyphen

0421ade

leseb merged commit 59ee496 into ceph:master Oct 31, 2016

wind0204 deleted the pr-monitor_interface_with_hyphens branch February 5, 2017 16:21

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

now you can use a network link name with a hyphen #1045

now you can use a network link name with a hyphen #1045

wind0204 commented Oct 24, 2016

leseb commented Oct 25, 2016

wind0204 commented Oct 25, 2016 •

edited

Loading

leseb commented Oct 25, 2016

wind0204 commented Oct 25, 2016 •

edited

Loading

leseb commented Oct 25, 2016

wind0204 commented Oct 25, 2016

leseb commented Oct 25, 2016

wind0204 commented Oct 26, 2016 •

edited

Loading

wind0204 commented Oct 26, 2016 •

edited

Loading

leseb commented Oct 31, 2016

leseb commented Oct 31, 2016

leseb commented Oct 31, 2016

font commented Oct 31, 2016

font commented Oct 31, 2016

leseb commented Oct 31, 2016

now you can use a network link name with a hyphen #1045

now you can use a network link name with a hyphen #1045

Conversation

wind0204 commented Oct 24, 2016

leseb commented Oct 25, 2016

wind0204 commented Oct 25, 2016 • edited Loading

leseb commented Oct 25, 2016

wind0204 commented Oct 25, 2016 • edited Loading

leseb commented Oct 25, 2016

wind0204 commented Oct 25, 2016

leseb commented Oct 25, 2016

wind0204 commented Oct 26, 2016 • edited Loading

wind0204 commented Oct 26, 2016 • edited Loading

leseb commented Oct 31, 2016

leseb commented Oct 31, 2016

leseb commented Oct 31, 2016

font commented Oct 31, 2016

font commented Oct 31, 2016

leseb commented Oct 31, 2016

wind0204 commented Oct 25, 2016 •

edited

Loading

wind0204 commented Oct 25, 2016 •

edited

Loading

wind0204 commented Oct 26, 2016 •

edited

Loading

wind0204 commented Oct 26, 2016 •

edited

Loading