[
https://issues.apache.org/jira/browse/HDDS-9665?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17787428#comment-17787428
]
Duong commented on HDDS-9665:
-----------------------------
[~xbis] I'm wondering if "dfsg" is linearly faster if you add more threads,
like "-t 80", in the large cluster.
Also, can you capture the following prometheus metric queries on the cluster
during the time tests run? The first one plots what how many OM requests
happening on OM side, and the second points out their latency.
{code:java}
increase(om_client_protocol_counter[30s]) / 30 {code}
{code:java}
increase(om_client_protocol_time[30s]) /
increase(om_client_protocol_counter[30s]) {code}
> DFS/ofs writes have very low performance
> ----------------------------------------
>
> Key: HDDS-9665
> URL: https://issues.apache.org/jira/browse/HDDS-9665
> Project: Apache Ozone
> Issue Type: Improvement
> Reporter: Christos Bisias
> Priority: Major
>
> This command is very slow
> {code:java}
> ozone freon dfsg -s10240000 -n10000 -t10 --path=ofs://om/pe-ofs/
> --prefix="ofs-100GB"{code}
>
> In contrast with this, which is fast
> {code:java}
> ozone freon ockg -s10240000 -n10000 -t10 -b pe-ofs -v vol1
> --prefix="ofs-100GB" {code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
