Mailing-List: contact issues-help@cloudstack.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@cloudstack.apache.org
Date: Sat, 15 Nov 2014 09:27:34 +0000 (UTC)
From: "Koushik Das (JIRA)" <jira@apache.org>
To: cloudstack-issues@incubator.apache.org
Message-ID: <JIRA.12747360.1412964532000.507944.1416043654053@Atlassian.JIRA>
In-Reply-To: <JIRA.12747360.1412964532000@Atlassian.JIRA>
References: <JIRA.12747360.1412964532000@Atlassian.JIRA>
 <JIRA.12747360.1412964532741@arcas>
Subject: [jira] [Commented] (CLOUDSTACK-7697) HA - No alerts being generated
 when SSVM/CPVM is being HA-ed to a different hosts.
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/CLOUDSTACK-7697?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14213477#comment-14213477 ] 

Koushik Das commented on CLOUDSTACK-7697:
-----------------------------------------

When host is detected as 'Down', all VMs running on that are scheduled for investigation (using HA worker threads). Based on investigation if VM is detected as alive, HA process is marked as completed. If it cannot be conclusively determined as alive/dead then the VM is fenced. After this if the VM is not HA enabled then the process is marked as completed. In this case no alert is generated for individual VMs as this is expected behaviour.

For SSVM/CPVM, there is a separate monitoring logic that will detect when there is no running SSVM/CPVM and spawn a new one. As part of this alerts will be generated.
In case of non-HA user VMs there is no need for alerts as the VM is not critical since it is not HA enabled. For HA enabled VMs it will tried to be started on another host and if there are any failures alerts are generated.

> HA - No alerts being generated when SSVM/CPVM is being HA-ed to a different hosts.
> ----------------------------------------------------------------------------------
>
>                 Key: CLOUDSTACK-7697
>                 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-7697
>             Project: CloudStack
>          Issue Type: Bug
>      Security Level: Public(Anyone can view this level - this is the default.) 
>          Components: Management Server
>    Affects Versions: 4.5.0
>         Environment: Build from 4.5
>            Reporter: Sangeetha Hariharan
>            Assignee: Koushik Das
>             Fix For: 4.5.0
>
>
> HA - No alerts being generated when SSVM/CPVM is being HA-ed to a different hosts.
> Steps to reproduce the problem:
> Zone with 1 cluster having 2 hosts.
> Bring down master host where SSVM and CPVM is running.
> All user Vms , SSVM and CPVM running in this host is HA-ed to another host.
> There is no Alert being generated for  SSVM and CPVM being detected as being stopped .
> Also there are no events/alerts being generated for all the user Vms that were detected as being stopped and started in a different host.
> Should we expect events/alerts being generated for these as well ?
> mysql> select * from alert;
> +----+--------------------------------------+------+------------+--------+----------------+---------------------------------------------------------------------------------------------+------------+---------------------+---------------------+----------+----------+--------------------+
> | id | uuid                                 | type | cluster_id | pod_id | data_center_id | subject                                                                                     | sent_count | created             | last_sent           | resolved | archived | name               |
> +----+--------------------------------------+------+------------+--------+----------------+---------------------------------------------------------------------------------------------+------------+---------------------+---------------------+----------+----------+--------------------+
> |  1 | aeef592e-3bb4-431e-911d-16280bf8a8ad |   14 |       NULL |      0 |              0 | Management network CIDR is not configured originally. Set it default to 10.223.130.0/24     |          1 | 2014-10-09 22:19:14 | 2014-10-09 22:19:14 | NULL     |        0 | ALERT.MANAGEMENT   |
> |  2 | 1a0bb67d-9346-4078-a80d-e6669116e7fd |   14 |       NULL |      0 |              0 | Management server node 10.223.130.101 is up                                                 |          1 | 2014-10-09 22:19:16 | 2014-10-09 22:19:16 | NULL     |        0 | ALERT.MANAGEMENT   |
> |  3 | 5c37924e-50cd-413f-a37a-ac275dbc46f9 |   13 |       NULL |      0 |              0 | No usage server process running                                                             |          1 | 2014-10-09 23:19:14 | 2014-10-09 23:19:14 | NULL     |        0 | ALERT.USAGE        |
> |  4 | 4d1b8b64-f59a-4405-a244-14e054297f04 |    2 |          1 |      1 |              1 | System Alert: Low Available Storage in cluster cluster1 pod pod1 of availability zone zone1 |          1 | 2014-10-09 23:39:44 | 2014-10-09 23:39:44 | NULL     |        0 | ALERT.STORAGE      |
> |  5 | aaf9bb96-799c-40d0-a652-96566c7ff47a |    7 |       NULL |      1 |              1 | Host is down, name: Rack3Host20.lab.vmops.com (id:1), availability zone: zone1, pod: pod1   |          1 | 2014-10-10 15:05:41 | 2014-10-10 15:05:41 | NULL     |        0 | ALERT.COMPUTE.HOST |
> +----+--------------------------------------+------+------------+--------+----------------+---------------------------------------------------------------------------------------------+------------+---------------------+---------------------+----------+----------+--------------------+
> 5 rows in set (0.00 sec)
> mysql>


--
This message was sent by Atlassian JIRA
(v6.3.4#6332)