Mailing-List: contact issues-help@cloudstack.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@cloudstack.apache.org
Date: Sat, 8 Nov 2014 08:08:33 +0000 (UTC)
From: "Sanjay Tripathi (JIRA)" <jira@apache.org>
To: cloudstack-issues@incubator.apache.org
Message-ID: <JIRA.12753901.1415434085000.452457.1415434113609@Atlassian.JIRA>
In-Reply-To: <JIRA.12753901.1415434085000@Atlassian.JIRA>
References: <JIRA.12753901.1415434085000@Atlassian.JIRA>
 <JIRA.12753901.1415434085903@arcas>
Subject: [jira] [Created] (CLOUDSTACK-7868) Failed
 storage.PrimaryStorageDownloadCommand leaves corrupt VDIs in primary
 storage
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit

Sanjay Tripathi created CLOUDSTACK-7868:
-------------------------------------------

             Summary: Failed storage.PrimaryStorageDownloadCommand leaves corrupt VDIs in primary storage
                 Key: CLOUDSTACK-7868
                 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-7868
             Project: CloudStack
          Issue Type: Bug
      Security Level: Public (Anyone can view this level - this is the default.)
         Environment: XS 6.0.2
            Reporter: Sanjay Tripathi
            Assignee: Sanjay Tripathi
            Priority: Critical
             Fix For: 4.5.0


The following situation happened:

1. Deploying VM from a template.
2. As part of the deploy CS would do a storage.PrimaryStorageDownloadCommand to get the vhd/template to primary storage.
3. If there is some type of MTU problem in the network, as a result the copying of the VHD will fail. VHD files in primary storage would end up with corrupt/missing header.
4. Since the copy command fails, CS tried to delete the vdi using VDI.destroy command
5. VDI.destroy fails because VDI file has missing/corrupt header
6. CS fails on downloading template and moves to another primary storage or cluster.

This process repeated until all XS clusters in the zone had a corrupt VDI in the storage. This causes future deploys (because they require sr-scan) to fail because the primary storages contain corrupt VDIs.


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