cloudstack-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Daan Hoogland <daan.hoogl...@gmail.com>
Subject Re: Change IP subnet of cluster [SOLVED]
Date Sat, 26 Jul 2014 05:13:55 GMT
H Carlos, glad you figured it out. A colleague had a similar issue but
his finding was that the host table included a timestamp to identify
the management server.

You are right about the replacement of the management server whether
it is mac or timestamp, this will pose a problem.

On Fri, Jul 25, 2014 at 10:43 PM, Carlos Reátegui <creategui@gmail.com> wrote:
> My system is back up and running.
>
> As I suspected in my second email the problem was related to the msid in the mshost table.
 Upon bringing up my system a new mshost entry was being created for the same MS and for some
reason it was unable to connect to my XenServer hosts.
>
> I decided to go back to my edited sql with the new IPs and change the existing mshost
entry to have the new msid value:
>
> sed -i.bak4 's/159090355471823/159090355471825/g' cloudstack_cloud-newips.sql
>
> I did a global replace since there are foreign key constraints on the msid that I saw
in the host and async_job tables
>
> After reloading this new sql and starting the MS everything is back to normal, but with
a new subnet for my hosts and guests (this is a basic network).
>
>
> Question for the Devs:
>
> Lets say the machine my MS was running on crashed and I replaced it with a new machine.
 Since the msid is derived from the MAC wouldn’t I have encountered this same problem and
not been able to have the new machine connect to the hosts?
>
> thanks,
> Carlos
>
>
>
>
> On Jul 25, 2014, at 8:46 AM, Carlos Reátegui <creategui@gmail.com> wrote:
>
>> Any thoughts out there?
>>
>> It keeps trying to connect to the hosts but it is unable to and there are no clues
in the logs as to why.  I am successfully connected with XenCenter to the pool and also am
able to ssh to all the hosts from the MS.
>>
>> What does “Disable Cluster” or “Unmanage Cluster” do?  Should I try that
and re-enable/manage?
>>
>> From the UI, things appear ok but starting any instance fails.
>>
>> Thanks,
>> Carlos
>>
>> Log snippet form this am:
>>
>> 2014-07-25 21:03:19,599 DEBUG [agent.manager.ClusteredAgentAttache] (StatsCollector-2:null)
Seq 2-65931749: Forwarding null to 159090355471823
>> 2014-07-25 21:03:19,599 DEBUG [agent.manager.ClusteredAgentAttache] (AgentManager-Handler-13:null)
Seq 2-65931749: Routing from 159090355471825
>> 2014-07-25 21:03:19,599 DEBUG [agent.manager.ClusteredAgentAttache] (AgentManager-Handler-13:null)
Seq 2-65931749: Link is closed
>> 2014-07-25 21:03:19,600 DEBUG [agent.manager.ClusteredAgentManagerImpl] (AgentManager-Handler-13:null)
Seq 2-65931749: MgmtId 159090355471825: Req: Resource [Host:2] is
>>  unreachable: Host 2: Link is closed
>> 2014-07-25 21:03:19,600 DEBUG [agent.manager.ClusteredAgentManagerImpl] (AgentManager-Handler-13:null)
Seq 2--1: MgmtId 159090355471825: Req: Routing to peer
>> 2014-07-25 21:03:19,601 DEBUG [agent.manager.ClusteredAgentManagerImpl] (AgentManager-Handler-14:null)
Seq 2--1: MgmtId 159090355471825: Req: Cancel request received
>> 2014-07-25 21:03:19,601 DEBUG [agent.manager.AgentAttache] (AgentManager-Handler-14:null)
Seq 2-65931749: Cancelling.
>> 2014-07-25 21:03:19,601 DEBUG [agent.manager.AgentAttache] (StatsCollector-2:null)
Seq 2-65931749: Waiting some more time because this is the current command
>> 2014-07-25 21:03:19,601 DEBUG [agent.manager.AgentAttache] (StatsCollector-2:null)
Seq 2-65931749: Waiting some more time because this is the current command
>> 2014-07-25 21:03:19,601 INFO  [utils.exception.CSExceptionErrorCode] (StatsCollector-2:null)
Could not find exception: com.cloud.exception.OperationTimedoutException in
>>  error code list for exceptions
>> 2014-07-25 21:03:19,601 WARN  [agent.manager.AgentAttache] (StatsCollector-2:null)
Seq 2-65931749: Timed out on null
>> 2014-07-25 21:03:19,601 DEBUG [agent.manager.AgentAttache] (StatsCollector-2:null)
Seq 2-65931749: Cancelling.
>> 2014-07-25 21:03:19,601 WARN  [agent.manager.AgentManagerImpl] (StatsCollector-2:null)
Operation timed out: Commands 65931749 to Host 2 timed out after 3600
>> 2014-07-25 21:03:19,601 WARN  [cloud.resource.ResourceManagerImpl] (StatsCollector-2:null)
Unable to obtain host 2 statistics.
>> 2014-07-25 21:03:19,601 WARN  [cloud.server.StatsCollector] (StatsCollector-2:null)
Received invalid host stats for host: 2
>> 2014-07-25 21:03:19,606 DEBUG [agent.manager.ClusteredAgentAttache] (StatsCollector-2:null)
Seq 3-602278373: Forwarding null to 159090355471823
>> 2014-07-25 21:03:19,607 DEBUG [agent.manager.ClusteredAgentAttache] (AgentManager-Handler-15:null)
Seq 3-602278373: Routing from 159090355471825
>> 2014-07-25 21:03:19,607 DEBUG [agent.manager.ClusteredAgentAttache] (AgentManager-Handler-15:null)
Seq 3-602278373: Link is closed
>> 2014-07-25 21:03:19,607 DEBUG [agent.manager.ClusteredAgentManagerImpl] (AgentManager-Handler-15:null)
Seq 3-602278373: MgmtId 159090355471825: Req: Resource [Host:3] is unreachable: Host 3: Link
is closed
>> 2014-07-25 21:03:19,608 DEBUG [agent.manager.ClusteredAgentManagerImpl] (AgentManager-Handler-15:null)
Seq 3--1: MgmtId 159090355471825: Req: Routing to peer
>> 2014-07-25 21:03:19,608 DEBUG [agent.manager.ClusteredAgentManagerImpl] (AgentManager-Handler-4:null)
Seq 3--1: MgmtId 159090355471825: Req: Cancel request received
>> 2014-07-25 21:03:19,609 DEBUG [agent.manager.AgentAttache] (AgentManager-Handler-4:null)
Seq 3-602278373: Cancelling.
>> 2014-07-25 21:03:19,609 DEBUG [agent.manager.AgentAttache] (StatsCollector-2:null)
Seq 3-602278373: Waiting some more time because this is the current command
>> 2014-07-25 21:03:19,609 DEBUG [agent.manager.AgentAttache] (StatsCollector-2:null)
Seq 3-602278373: Waiting some more time because this is the current command
>> 2014-07-25 21:03:19,609 INFO  [utils.exception.CSExceptionErrorCode] (StatsCollector-2:null)
Could not find exception: com.cloud.exception.OperationTimedoutException in error code list
for exceptions
>> 2014-07-25 21:03:19,609 WARN  [agent.manager.AgentAttache] (StatsCollector-2:null)
Seq 3-602278373: Timed out on null
>> 2014-07-25 21:03:19,609 DEBUG [agent.manager.AgentAttache] (StatsCollector-2:null)
Seq 3-602278373: Cancelling.
>> 2014-07-25 21:03:19,609 WARN  [agent.manager.AgentManagerImpl] (StatsCollector-2:null)
Operation timed out: Commands 602278373 to Host 3 timed out after 3600
>> 2014-07-25 21:03:19,609 WARN  [cloud.resource.ResourceManagerImpl] (StatsCollector-2:null)
Unable to obtain host 3 statistics.
>> 2014-07-25 21:03:19,609 WARN  [cloud.server.StatsCollector] (StatsCollector-2:null)
Received invalid host stats for host: 3
>> 2014-07-25 21:03:19,614 DEBUG [agent.manager.ClusteredAgentAttache] (StatsCollector-2:null)
Seq 5-1311574501: Forwarding null to 159090355471823
>> 2014-07-25 21:03:19,617 DEBUG [agent.manager.ClusteredAgentAttache] (AgentManager-Handler-1:null)
Seq 5-1311574501: Routing from 159090355471825
>> 2014-07-25 21:03:19,617 DEBUG [agent.manager.ClusteredAgentAttache] (AgentManager-Handler-1:null)
Seq 5-1311574501: Link is closed
>> 2014-07-25 21:03:19,617 DEBUG [agent.manager.ClusteredAgentManagerImpl] (AgentManager-Handler-1:null)
Seq 5-1311574501: MgmtId 159090355471825: Req: Resource [Host:5] is unreachable: Host 5: Link
is closed
>>
>>
>>
>> On Jul 24, 2014, at 10:59 PM, Carlos Reátegui <carlos@reategui.com> wrote:
>>
>>> Not sure if it is related but I see 2 entries in the mshost for my same server
but with different msid.  Both show as ‘Up’.  In reading the table comments it seems the
msid is based on the MAC.  I am guessing this may be due to using a bond and that it may be
have selected a different NIC to get the bond MAC from.  Is it ok to have both of these entries?
 Should I mark the old one as Down?
>>>
>>> Along these lines is there something similar with the hosts and that is why the
MS is having problems connecting to them, ie. the MACs don’t match?
>>>
>>> thanks,
>>> Carlos
>>>
>>>
>>> On Jul 24, 2014, at 3:35 PM, Carlos Reategui <carlos@reategui.com> wrote:
>>>
>>>> Hi All,
>>>>
>>>> Had to move one of my clusters to a new subnet but it is not working (e.g.
192.168.1.0/24 to 10.100.1.0/24).  These are the steps I took:
>>>>
>>>> Environment: CS 4.1.1 on Ubuntu 12.04, XenServer 6.1, Shared NFS SR.
>>>>
>>>> 1) stopped all instances using cloudstack UI
>>>> 2) stop cloudstack-management service on MS
>>>> 3) Used XenCenter to kill the system VMs (no other instances running)
>>>> 4) Created backup of cloud db.
>>>> 5) Followed http://support.citrix.com/article/CTX123477 and successfully
changed the IP of hosts.  According to XenCenter everything is good including SR.
>>>> 6) Changed IP of MS
>>>> 7) verified communication between MS and Hosts using ssh and ping with new
IPs.
>>>> 8) used sed to search and replace all old IPs with new IPs in cloud backup
sql file (e.g. sed -i.bak 's/192.168.1./10.100.1./g' clouddb.sql).
>>>> 9) visually verified all diffs in the sql file and made sure no references
to 192.168 left.
>>>> 10) loaded up new sql
>>>> 11) search all files under /etc on MS for old IP. found and edited: /etc/cloudstack/management/db.properties
>>>> 12) start cloudstack-management service on MS
>>>>
>>>> Unfortunately things are not working.  The MS is apparently unable to connect
to the hosts but I can not figure out why from the logs.
>>>>
>>>> Logs here: https://www.dropbox.com/s/s5glxrbyatmsoug/management-server.log
>>>>
>>>> Any help recovering is appreciated.  I do not want to have to re-install
and create/import template for each of the instance VHDs.
>>>>
>>>> thank you,
>>>> -Carlos
>>>
>>
>



-- 
Daan

Mime
View raw message