ERROR state after server stop

classic Classic list List threaded Threaded
9 messages Options
Reply | Threaded
Open this post in threaded view
|

ERROR state after server stop

jcalcote
Hi Yaguang,

I have a server in an ERROR state and don't know where to go from here.

I have a three-server configuration: parkcity is my controller, alta and brighton are my OSD servers, all three hosts are also ceph monitors. VSM was running fine - the system was healthy. I decided to try VSM | Manage Servers | Cluster Server List | Stop Servers. I selected one of my OSD servers (alta). I got a spinner that disappeared briefly for a split-second and then reappeared and ran for several minutes. I refreshed the screen and went back to the main VSM | Manage Servers screen. I now see that the Status field for that server is ERROR. How can I tell what's going on here.

Here's some relevant information found in the vsm-api.log file around the time the "stop server" command was executed:

2015-12-17 12:33:24     INFO [vsm.api.v1.servers] DEBUG stop-server body {u'servers': [{u'remove_storage': False, u'cluster_id': 1, u'id': u'2', u'remove_monitor': True}]}
2015-12-17 12:33:25    ERROR [vsm.api.openstack.wsgi] Exception handling resource: cannot concatenate 'str' and 'NoneType' objects
Traceback (most recent call last):

  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/amqp.py", line 415, in _process_data
    rval = self.proxy.dispatch(ctxt, version, method, **args)

  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/dispatcher.py", line 130, in dispatch
    return getattr(proxyobj, method)(ctxt, **kwargs)

  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/manager.py", line 835, in stop_server
    return self.ceph_driver.stop_server(context, node_id)

  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/driver.py", line 1557, in stop_server
    cephconfigparser.CephConfigParser(FLAGS.ceph_conf)

  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/cephconfigparser.py", line 255, in __init__
    self._load_ceph_conf_from_db()

  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/cephconfigparser.py", line 247, in _load_ceph_conf_from_db
    line = line + ' ' + file_system

Thanks,
John
Reply | Threaded
Open this post in threaded view
|

RE: ERROR state after server stop

ywang19
Administrator

Is this thread(http://vsm-discuss.33411.n7.nabble.com/Parser-Bug-in-cephconfigparser-py-td280.html) related to your case?

 

 

From: jcalcote [via vsm-discuss] [mailto:ml-node+[hidden email]]
Sent: Friday, December 18, 2015 4:01 AM
To: Wang, Yaguang
Subject: ERROR state after server stop

 

Hi Yaguang,

I have a server in an ERROR state and don't know where to go from here.

I have a three-server configuration: parkcity is my controller, alta and brighton are my OSD servers, all three hosts are also ceph monitors. VSM was running fine - the system was healthy. I decided to try VSM | Manage Servers | Cluster Server List | Stop Servers. I selected one of my OSD servers (alta). I got a spinner that disappeared briefly for a split-second and then reappeared and ran for several minutes. I refreshed the screen and went back to the main VSM | Manage Servers screen. I now see that the Status field for that server is ERROR. How can I tell what's going on here.

Here's some relevant information found in the vsm-api.log file around the time the "stop server" command was executed:

 
2015-12-17 12:33:24     INFO [vsm.api.v1.servers] DEBUG stop-server body {u'servers': [{u'remove_storage': False, u'cluster_id': 1, u'id': u'2', u'remove_monitor': True}]}
2015-12-17 12:33:25    ERROR [vsm.api.openstack.wsgi] Exception handling resource: cannot concatenate 'str' and 'NoneType' objects
Traceback (most recent call last):
 
  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/amqp.py", line 415, in _process_data
    rval = self.proxy.dispatch(ctxt, version, method, **args)
 
  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/dispatcher.py", line 130, in dispatch
    return getattr(proxyobj, method)(ctxt, **kwargs)
 
  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/manager.py", line 835, in stop_server
    return self.ceph_driver.stop_server(context, node_id)
 
  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/driver.py", line 1557, in stop_server
    cephconfigparser.CephConfigParser(FLAGS.ceph_conf)
 
  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/cephconfigparser.py", line 255, in __init__
    self._load_ceph_conf_from_db()
 
  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/cephconfigparser.py", line 247, in _load_ceph_conf_from_db
    line = line + ' ' + file_system


Thanks,
John


If you reply to this email, your message will be added to the discussion below:

http://vsm-discuss.33411.n7.nabble.com/ERROR-state-after-server-stop-tp307.html

To start a new topic under vsm-discuss, email [hidden email]
To unsubscribe from vsm-discuss, click here.
NAML

Reply | Threaded
Open this post in threaded view
|

RE: ERROR state after server stop

jcalcote
I don't think so - our ceph.conf has no lines with multiple equal signs, however, our ceph.conf does have a lot of whitespace in it - that is every line has white space on both sides of the equal sign - just in case that makes a difference.

John
Reply | Threaded
Open this post in threaded view
|

Re: ERROR state after server stop

jcalcote
In reply to this post by jcalcote
I believe this issue is more likely related to your comments on this thread Yaguang:

http://vsm-discuss.33411.n7.nabble.com/Expanding-an-imported-cluster-via-native-methods-tp326p327.html

where you mention that you've found issues with "stop server" in vsm when cluster is deployed with ceph-deploy.

Do you have a solution for these issues yet? You mentioned also that you have a new strategy for stop server - has this been integrated yet? I can pull and update to get these changes if they're available.

Thanks,
John
Reply | Threaded
Open this post in threaded view
|

RE: ERROR state after server stop

ywang19
Administrator

This is exactly what we are flighting, as clusters deployed through ceph-deploy uses some different ceph processes launching approaches.

 

 

 

From: jcalcote [via vsm-discuss] [mailto:ml-node+[hidden email]]
Sent: Tuesday, December 29, 2015 12:14 AM
To: Wang, Yaguang
Subject: Re: ERROR state after server stop

 

I believe this issue is more likely related to your comments on this thread Yaguang:

http://vsm-discuss.33411.n7.nabble.com/Expanding-an-imported-cluster-via-native-methods-tp326p327.html

where you mention that you've found issues with "stop server" in vsm when cluster is deployed with ceph-deploy.

Do you have a solution for these issues yet? You mentioned also that you have a new strategy for stop server - has this been integrated yet? I can pull and update to get these changes if they're available.

Thanks,
John


If you reply to this email, your message will be added to the discussion below:

http://vsm-discuss.33411.n7.nabble.com/ERROR-state-after-server-stop-tp307p333.html

To start a new topic under vsm-discuss, email [hidden email]
To unsubscribe from vsm-discuss, click here.
NAML

Reply | Threaded
Open this post in threaded view
|

RE: ERROR state after server stop

jcalcote
Hi Yaguang,

I've updated to the 2.1 beta release and am still experiencing this issue - I have the same traceback (as above) in the vsm-scheduler.log.

TypeError: cannot concatenate 'str' and 'NoneType' objects

2016-01-02 23:40:46    ERROR [vsm.openstack.common.rpc.common] Returning exception cannot concatenate 'str' and 'NoneType' objects
Traceback (most recent call last):

  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/amqp.py", line 415, in _process_data
    rval = self.proxy.dispatch(ctxt, version, method, **args)

  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/dispatcher.py", line 130, in dispatch
    return getattr(proxyobj, method)(ctxt, **kwargs)

  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/manager.py", line 836, in stop_server
    return self.ceph_driver.stop_server(context, node_id)

  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/driver.py", line 1627, in stop_server
    cephconfigparser.CephConfigParser(FLAGS.ceph_conf)

  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/cephconfigparser.py", line 251, in __init__
    self._load_ceph_conf_from_db()

  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/cephconfigparser.py", line 243, in _load_ceph_conf_from_db
    line = line + ' ' + file_system

TypeError: cannot concatenate 'str' and 'NoneType' objects
 to caller
2016-01-02 23:40:46    ERROR [vsm.openstack.common.rpc.common] ['Traceback (most recent call last):\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/amqp.py", line 415, in _process_data\n    rval = self.proxy.dispatch(ctxt, version, method, **args)\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/dispatcher.py", line 130, in dispatch\n    return getattr(proxyobj, method)(ctxt, **kwargs)\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/utils.py", line 1540, in _deco\n    ret = func(*args, **kwargs)\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/scheduler/manager.py", line 910, in stop_server\n    res[\'host\'])\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/rpcapi.py", line 286, in stop_server\n    version=\'1.0\', timeout=6000)\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/proxy.py", line 80, in call\n    return rpc.call(context, self._get_topic(topic), msg, timeout, need_try=need_try)\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/__init__.py", line 145, in call\n    ret = __call(context, topic, msg, timeout, check_for_lock)\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/__init__.py", line 135, in __call\n    return _get_impl().call(CONF, context, topic, msg, timeout)\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/impl_kombu.py", line 784, in call\n    rpc_amqp.get_connection_pool(conf, Connection))\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/amqp.py", line 592, in call\n    rv = list(rv)\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/amqp.py", line 545, in __iter__\n    raise result\n', 'TypeError: cannot concatenate \'str\' and \'NoneType\' objects\nTraceback (most recent call last):\n\n  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/amqp.py", line 415, in _process_data\n    rval = self.proxy.dispatch(ctxt, version, method, **args)\n\n  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/dispatcher.py", line 130, in dispatch\n    return getattr(proxyobj, method)(ctxt, **kwargs)\n\n  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/manager.py", line 836, in stop_server\n    return self.ceph_driver.stop_server(context, node_id)\n\n  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/driver.py", line 1627, in stop_server\n    cephconfigparser.CephConfigParser(FLAGS.ceph_conf)\n\n  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/cephconfigparser.py", line 251, in __init__\n    self._load_ceph_conf_from_db()\n\n  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/cephconfigparser.py", line 243, in _load_ceph_conf_from_db\n    line = line + \' \' + file_system\n\nTypeError: cannot concatenate \'str\' and \'NoneType\' objects\n\n']

This was found in the vsm-api.log:

2016-01-02 23:37:46     INFO [vsm.api.v1.servers] DEBUG stop-server body {u'servers': [{u'remove_storage': False, u'cluster_id': 1, u'id': u'3', u'remove_monitor': True}]}
2016-01-02 23:37:47    ERROR [vsm.api.openstack.wsgi] Exception handling resource: cannot concatenate 'str' and 'NoneType' objects
Traceback (most recent call last):

  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/amqp.py", line 415, in _process_data
    rval = self.proxy.dispatch(ctxt, version, method, **args)

  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/dispatcher.py", line 130, in dispatch
    return getattr(proxyobj, method)(ctxt, **kwargs)

  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/manager.py", line 836, in stop_server
    return self.ceph_driver.stop_server(context, node_id)

  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/driver.py", line 1627, in stop_server
    cephconfigparser.CephConfigParser(FLAGS.ceph_conf)

  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/cephconfigparser.py", line 251, in __init__
    self._load_ceph_conf_from_db()

  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/cephconfigparser.py", line 243, in _load_ceph_conf_from_db
    line = line + ' ' + file_system

TypeError: cannot concatenate 'str' and 'NoneType' objects


It appears there is still an issue with stop server. Can you please help me figure out where I might look to give you more information on this issue?

Thanks,
John
Reply | Threaded
Open this post in threaded view
|

RE: ERROR state after server stop

ywang19
Administrator

Hi John,

 

Could you give out the operation path? We will try to reproduce it locally to identify the root cause.

 

-yaguang

 

From: jcalcote [via vsm-discuss] [mailto:ml-node+[hidden email]]
Sent: Sunday, January 03, 2016 2:04 PM
To: Wang, Yaguang
Subject: RE: ERROR state after server stop

 

Hi Yaguang,

I've updated to the 2.1 beta release and am still experiencing this issue - I have the same traceback (as above) in the vsm-scheduler.log.

 
TypeError: cannot concatenate 'str' and 'NoneType' objects
 
2016-01-02 23:40:46    ERROR [vsm.openstack.common.rpc.common] Returning exception cannot concatenate 'str' and 'NoneType' objects
Traceback (most recent call last):
 
  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/amqp.py", line 415, in _process_data
    rval = self.proxy.dispatch(ctxt, version, method, **args)
 
  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/dispatcher.py", line 130, in dispatch
    return getattr(proxyobj, method)(ctxt, **kwargs)
 
  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/manager.py", line 836, in stop_server
    return self.ceph_driver.stop_server(context, node_id)
 
  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/driver.py", line 1627, in stop_server
    cephconfigparser.CephConfigParser(FLAGS.ceph_conf)
 
  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/cephconfigparser.py", line 251, in __init__
    self._load_ceph_conf_from_db()
 
  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/cephconfigparser.py", line 243, in _load_ceph_conf_from_db
    line = line + ' ' + file_system
 
TypeError: cannot concatenate 'str' and 'NoneType' objects
 to caller
2016-01-02 23:40:46    ERROR [vsm.openstack.common.rpc.common] ['Traceback (most recent call last):\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/amqp.py", line 415, in _process_data\n    rval = self.proxy.dispatch(ctxt, version, method, **args)\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/dispatcher.py", line 130, in dispatch\n    return getattr(proxyobj, method)(ctxt, **kwargs)\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/utils.py", line 1540, in _deco\n    ret = func(*args, **kwargs)\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/scheduler/manager.py", line 910, in stop_server\n    res[\'host\'])\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/rpcapi.py", line 286, in stop_server\n    version=\'1.0\', timeout=6000)\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/proxy.py", line 80, in call\n    return rpc.call(context, self._get_topic(topic), msg, timeout, need_try=need_try)\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/__init__.py", line 145, in call\n    ret = __call(context, topic, msg, timeout, check_for_lock)\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/__init__.py", line 135, in __call\n    return _get_impl().call(CONF, context, topic, msg, timeout)\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/impl_kombu.py", line 784, in call\n    rpc_amqp.get_connection_pool(conf, Connection))\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/amqp.py", line 592, in call\n    rv = list(rv)\n', '  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/amqp.py", line 545, in __iter__\n    raise result\n', 'TypeError: cannot concatenate \'str\' and \'NoneType\' objects\nTraceback (most recent call last):\n\n  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/amqp.py", line 415, in _process_data\n    rval = self.proxy.dispatch(ctxt, version, method, **args)\n\n  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/dispatcher.py", line 130, in dispatch\n    return getattr(proxyobj, method)(ctxt, **kwargs)\n\n  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/manager.py", line 836, in stop_server\n    return self.ceph_driver.stop_server(context, node_id)\n\n  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/driver.py", line 1627, in stop_server\n    cephconfigparser.CephConfigParser(FLAGS.ceph_conf)\n\n  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/cephconfigparser.py", line 251, in __init__\n    self._load_ceph_conf_from_db()\n\n  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/cephconfigparser.py", line 243, in _load_ceph_conf_from_db\n    line = line + \' \' + file_system\n\nTypeError: cannot concatenate \'str\' and \'NoneType\' objects\n\n']


This was found in the vsm-api.log:

 
2016-01-02 23:37:46     INFO [vsm.api.v1.servers] DEBUG stop-server body {u'servers': [{u'remove_storage': False, u'cluster_id': 1, u'id': u'3', u'remove_monitor': True}]}
2016-01-02 23:37:47    ERROR [vsm.api.openstack.wsgi] Exception handling resource: cannot concatenate 'str' and 'NoneType' objects
Traceback (most recent call last):
 
  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/amqp.py", line 415, in _process_data
    rval = self.proxy.dispatch(ctxt, version, method, **args)
 
  File "/usr/local/lib/python2.7/dist-packages/vsm/openstack/common/rpc/dispatcher.py", line 130, in dispatch
    return getattr(proxyobj, method)(ctxt, **kwargs)
 
  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/manager.py", line 836, in stop_server
    return self.ceph_driver.stop_server(context, node_id)
 
  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/driver.py", line 1627, in stop_server
    cephconfigparser.CephConfigParser(FLAGS.ceph_conf)
 
  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/cephconfigparser.py", line 251, in __init__
    self._load_ceph_conf_from_db()
 
  File "/usr/local/lib/python2.7/dist-packages/vsm/agent/cephconfigparser.py", line 243, in _load_ceph_conf_from_db
    line = line + ' ' + file_system
 
TypeError: cannot concatenate 'str' and 'NoneType' objects
 


It appears there is still an issue with stop server. Can you please help me figure out where I might look to give you more information on this issue?

Thanks,
John


If you reply to this email, your message will be added to the discussion below:

http://vsm-discuss.33411.n7.nabble.com/ERROR-state-after-server-stop-tp307p337.html

To start a new topic under vsm-discuss, email [hidden email]
To unsubscribe from vsm-discuss, click here.
NAML

Reply | Threaded
Open this post in threaded view
|

RE: ERROR state after server stop

jcalcote
Hi Yaguang,

I've discovered the root cause of the issue. As the stacktrace shows, the exception is being thrown from line 243 of vsm/agent/cephconfigparser.py:

    def _load_ceph_conf_from_db(self):
        ...
        parser = Parser()
        parser.read(FLAGS.ceph_conf)
        fs_type = parser.get('osd', 'osd mkfs type')
        mount_attr = parser.get('osd', 'osd mount options %s' % fs_type)
        file_system = parser.get('osd', 'osd mkfs type')

        for sec in parser.sections():
            if sec.find('osd.') != -1:
                osd_id = sec.split('.')[1]
                mount_path = os.path.join(FLAGS.osd_data_path, "osd%s" % osd_id)
                mount_disk = parser.get(sec, 'devs')
                mount_host = parser.get(sec, 'host')
                if FLAGS.host == mount_host:
                    line = mount_disk + ' ' + mount_path
243 =>              line = line + ' ' + file_system
                    line = line + ' ' + mount_attr + ' 0 0'
                    line = line + ' ' + '## forvsmosd'
                    utils.write_file_as_root('/etc/fstab', line)

Ultimately, the problem is that our ceph.conf file doesn't have the following lines in the [osd] section:

[osd]
...
osd mkfs type = xfs
osd mount options xfs = rw,noatime,inode64,logbsize=256k,delaylog
...

We also don't have these other [osd] section entries that a standard VSM deployment has:

[osd]
...
osd crush update on start = false
filestore xattr use omap = true
keyring = /etc/ceph/keyring.$name
osd data = /var/lib/ceph/osd/osd$id
osd heartbeat grace = 10
osd heartbeat interval = 10
...

I believe we can add the first set without issue, but can you tell me if VSM will care if the others are missing?

One final question: Our ceph.conf file uses the underscore form of the attribute names - that is, we use (for example) "osd_journal_size" instead of "osd journal size". Does VSM's ceph.conf file parser understand both forms, or do we need to ensure we are using the space form?

Thanks,
John
Reply | Threaded
Open this post in threaded view
|

RE: ERROR state after server stop

ywang19
Administrator

For the missing [osd] sections in ceph.conf, yes, it’s required. I’m working on a functionalities to automatic fixup those missing sections for those clusters deployed by ceph-deploy, still needs a few days to get it ready.

 

For underscore connected parameter names, my recent commit for cephconfigparser.py should support it.

 

 

 

 

From: jcalcote [via vsm-discuss] [mailto:ml-node+[hidden email]]
Sent: Tuesday, January 05, 2016 1:39 AM
To: Wang, Yaguang
Subject: RE: ERROR state after server stop

 

Hi Yaguang,

I've discovered the root cause of the issue. As the stacktrace shows, the exception is being thrown from line 243 of vsm/agent/cephconfigparser.py:

 
    def _load_ceph_conf_from_db(self):
        ...
        parser = Parser()
        parser.read(FLAGS.ceph_conf)
        fs_type = parser.get('osd', 'osd mkfs type')
        mount_attr = parser.get('osd', 'osd mount options %s' % fs_type)
        file_system = parser.get('osd', 'osd mkfs type')
 
        for sec in parser.sections():
            if sec.find('osd.') != -1:
                osd_id = sec.split('.')[1]
                mount_path = os.path.join(FLAGS.osd_data_path, "osd%s" % osd_id)
                mount_disk = parser.get(sec, 'devs')
                mount_host = parser.get(sec, 'host')
                if FLAGS.host == mount_host:
                    line = mount_disk + ' ' + mount_path
243 =>              line = line + ' ' + file_system
                    line = line + ' ' + mount_attr + ' 0 0'
                    line = line + ' ' + '## forvsmosd'
                    utils.write_file_as_root('/etc/fstab', line)


Ultimately, the problem is that our ceph.conf file doesn't have the following lines in the [osd] section:

 
[osd]
...
osd mkfs type = xfs
osd mount options xfs = rw,noatime,inode64,logbsize=256k,delaylog
...


We also don't have these other [osd] section entries that a standard VSM deployment has:

 
[osd]
...
osd crush update on start = false
filestore xattr use omap = true
keyring = /etc/ceph/keyring.$name
osd data = /var/lib/ceph/osd/osd$id
osd heartbeat grace = 10
osd heartbeat interval = 10
...


I believe we can add the first set without issue, but can you tell me if VSM will care if the others are missing?

One final question: Our ceph.conf file uses the underscore form of the attribute names - that is, we use (for example) "osd_journal_size" instead of "osd journal size". Does VSM's ceph.conf file parser understand both forms, or do we need to ensure we are using the space form?

Thanks,
John


If you reply to this email, your message will be added to the discussion below:

http://vsm-discuss.33411.n7.nabble.com/ERROR-state-after-server-stop-tp307p342.html

To start a new topic under vsm-discuss, email [hidden email]
To unsubscribe from vsm-discuss, click here.
NAML