[
https://issues.apache.org/jira/browse/HDDS-5560?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17640889#comment-17640889
]
Ethan Rose commented on HDDS-5560:
----------------------------------
There's a bit of a challenge with fixing this. The cluster ID comes from SCM
response in the VersionEndpointTask. At this point the DN has already
initialized and written/read its VERSION file. Replacing the cluster ID field
with the DN ID might be better since we have that value at startup and no one
is reading this value. Need to check downgrade implications though.
> Datanodes' dnlayoutversion/VERSION file contains incorrect cluster ID
> ---------------------------------------------------------------------
>
> Key: HDDS-5560
> URL: https://issues.apache.org/jira/browse/HDDS-5560
> Project: Apache Ozone
> Issue Type: Sub-task
> Reporter: Ethan Rose
> Assignee: Nishit Patira
> Priority: Major
> Labels: newbie
>
> Taken from a datanode as part of a docker cluster
> Global version file at metadata/dnlayoutversion/VERSION (incorrect cluster
> ID):
> {code:java}
> #Fri Jul 23 17:52:54 UTC 2021
> firstUpgradeActionLayoutVersion=2
> cTime=1627062428170
> clusterID=18539b4c-28ae-4dfb-a668-6437c2df56f6
> nodeType=DATANODE
> layoutVersion=2
> {code}
> Volume specific version file at hdds/hdds/VERSION (correct):
> {code:java}
> #Fri Jul 23 17:45:20 UTC 2021
> datanodeUuid=18539b4c-28ae-4dfb-a668-6437c2df56f6
> ctime=1627062320832
> clusterID=CID-faab32bd-91e1-4ffc-a942-e4252d8af722
> layOutVersion=1
> storageID=DS-1876fa77-6c75-4b91-b724-7802a362af52
> {code}
> Global datanode.id file (correct):
> {code:java}
> !!org.apache.hadoop.ozone.container.common.helpers.DatanodeIdYaml$DatanodeDetailsYaml
> {
> certSerialId: null,
> currentVersion: 1,
> hostName: b08914e5c4ae,
> initialVersion: 1,
> ipAddress: 10.9.0.15,
> persistedOpState: IN_SERVICE,
> persistedOpStateExpiryEpochSec: 0,
> portDetails: {
> },
> uuid: 18539b4c-28ae-4dfb-a668-6437c2df56f6
> }
> {code}
>
> All datanodes use the same cluster ID in their volume specific version files,
> and SCM shows a matching cluster ID. This means the value
> CID-faab32bd-91e1-4ffc-a942-e4252d8af722 is the correct cluster ID. The other
> two files show the datanode UUID as 18539b4c-28ae-4dfb-a668-6437c2df56f6,
> which is incorrectly used as the cluster ID in the global datanode version
> file.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
