You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ozone.apache.org by "Siyao Meng (Jira)" <ji...@apache.org> on 2020/02/20 20:28:00 UTC
[jira] [Updated] (HDDS-2929) Implement ofs://: temp directory mount
[ https://issues.apache.org/jira/browse/HDDS-2929?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Siyao Meng updated HDDS-2929:
-----------------------------
Description:
Because of ofs:// filesystem hierarchy starts with volume then bucket, an application typically won't be able to write directly under a first-level folder, e.g. ofs://service-id1/tmp/. /tmp/ is a special case that we need to handle since that is the default location most legacy Hadoop applications write to for swap/temporary files. In order to address this, we would introduce /tmp/ as a client-side "mount" to another Ozone bucket.
Note that the preliminary implementation would only allow for /tmp/ to be a mount but not any user-defined path.
This depends on HDDS-2840 and HDDS-2928.
Demo PR on my fork: https://github.com/smengcl/hadoop-ozone/pull/1
was:
Because of ofs:// filesystem hierarchy starts with volume then bucket, an application typically won't be able to write directly under a first-level folder, e.g. ofs://service-id1/tmp/. /tmp/ is a special case that we need to handle since that is the default location most legacy Hadoop applications write to for swap/temporary files. In order to address this, we would introduce /tmp/ as a client-side "mount" to another Ozone bucket.
Note that the preliminary implementation would only allow for /tmp/ to be a mount but not any user-defined path.
This depends on HDDS-2840.
> Implement ofs://: temp directory mount
> --------------------------------------
>
> Key: HDDS-2929
> URL: https://issues.apache.org/jira/browse/HDDS-2929
> Project: Hadoop Distributed Data Store
> Issue Type: Sub-task
> Reporter: Siyao Meng
> Assignee: Siyao Meng
> Priority: Major
>
> Because of ofs:// filesystem hierarchy starts with volume then bucket, an application typically won't be able to write directly under a first-level folder, e.g. ofs://service-id1/tmp/. /tmp/ is a special case that we need to handle since that is the default location most legacy Hadoop applications write to for swap/temporary files. In order to address this, we would introduce /tmp/ as a client-side "mount" to another Ozone bucket.
> Note that the preliminary implementation would only allow for /tmp/ to be a mount but not any user-defined path.
> This depends on HDDS-2840 and HDDS-2928.
> Demo PR on my fork: https://github.com/smengcl/hadoop-ozone/pull/1
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: ozone-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: ozone-issues-help@hadoop.apache.org