[ceph-users] ceph-iscsi upgrade issue

Mike Christie mchristi at redhat.com
Wed Oct 10 13:42:27 PDT 2018


On 10/10/2018 03:18 PM, Steven Vacaroaia wrote:
> so, it seems OSD03 is having issues when creating disks ( I can create
> target and hosts ) - here is an excerpt from api.log
> Please note I can create disk on the other node  
> 
> 2018-10-10 16:03:03,369    DEBUG [lun.py:381:allocate()] - LUN.allocate
> starting, listing rbd devices
> 2018-10-10 16:03:03,381    DEBUG [lun.py:384:allocate()] - rados pool
> 'rbd' contains the following - [u'datastoretest', u'datastoretest2',
> u'disk1', u'dstest', u'scstimage1', u'test2']
> 2018-10-10 16:03:03,382    DEBUG [lun.py:389:allocate()] - Hostname
> Check - this host is osd03, target host for allocations is osd03
> 2018-10-10 16:03:03,428    DEBUG [common.py:272:add_item()] -
> (Config.add_item) config updated to {u'updated': u'2018/10/10 20:02:24',
> u'disks': {'rbd.dstest2': {'created': '2018/10/10 20:03:03'}},
> u'created': u'2018/10/10 19:56:05', u'clients': {}, u'epoch': 4,
> u'version': 3, u'gateways': {u'osd03': {u'gateway_ip_list':
> [u'10.10.30.183', u'10.10.30.181'], u'active_luns': 0, u'created':
> u'2018/10/10 20:02:12', u'updated': u'2018/10/10 20:02:24', u'iqn':
> u'iqn.2003-01.com.redhat.iscsi-gw:iscsi-ceph', u'inactive_portal_ips':
> [u'10.10.30.181'], u'portal_ip_address': u'10.10.30.183', u'tpgs': 2},
> u'osd01': {u'gateway_ip_list': [u'10.10.30.183', u'10.10.30.181'],
> u'active_luns': 0, u'created': u'2018/10/10 20:02:24', u'updated':
> u'2018/10/10 20:02:24', u'iqn':
> u'iqn.2003-01.com.redhat.iscsi-gw:iscsi-ceph', u'inactive_portal_ips':
> [u'10.10.30.183'], u'portal_ip_address': u'10.10.30.181', u'tpgs': 2},
> u'iqn': u'iqn.2003-01.com.redhat.iscsi-gw:iscsi-ceph', u'ip_list':
> [u'10.10.30.183', u'10.10.30.181'], u'created': u'2018/10/10 20:01:39'},
> u'groups': {}}
> 2018-10-10 16:03:03,429     INFO [lun.py:405:allocate()] -
> (LUN.allocate) created rbd/dstest2 successfully
> 2018-10-10 16:03:03,429    DEBUG [lun.py:444:allocate()] - Check the rbd
> image size matches the request
> 2018-10-10 16:03:03,429    DEBUG [lun.py:467:allocate()] - Begin
> processing LIO mapping
> 2018-10-10 16:03:03,429     INFO [lun.py:656:add_dev_to_lio()] -
> (LUN.add_dev_to_lio) Adding image 'rbd.dstest2' to LIO
> 2018-10-10 16:03:03,429    *DEBUG [lun.py:666:add_dev_to_lio()] -
> control="max_data_area_mb=8"*



> python2-rtslib-2.1.fb67-3.fc28.noarch.rpm

Did you get this from the fedora repo? If so that is going to cause
problems, because it is missing some support we need. I think we should
have got an error after the message above though, but maybe that it is
getting logged to a different file.

You should remove it and use the version with came with ceph-iscsi-*. If
you got them from https://shaman.ceph.com/repos/ then use the rtslib
from there. If you got the ceph-iscsi-* rpms from RHCS then use the
rtslib from there.

You could also use the upstream github one if you are building them
yourself.


More information about the ceph-users mailing list