Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Nfsdist tool #1347

Merged
merged 6 commits into from
Sep 12, 2017
Merged

Nfsdist tool #1347

merged 6 commits into from
Sep 12, 2017

Conversation

samuelnair
Copy link
Contributor

This is a pull request to add the nfsdist tool, which is to compliment the nfsslower tool that i had added in earlier and in keeping with the suite of {etx4,xfs,zfs}(slower|dist) tools that we currently have.

As before, following the CONTRIBUTING-SCRIPTS.md guidelines and added in the smoke test as well.

@goldshtn
Copy link
Collaborator

[buildbot, ok to test]

8191 usecs. Write traffic is largely clusstered in the 64 - 127 usecs bracket.
The faster read traffic is probably coming from a filesystem cache and the slower
traffic from disk. The reason why the writes are so consistenly fast is becuase
this example test was run on a couple of VM's and i believe the hypervisor was
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Capital "I".

26K reads falling within 8 - 30 usecs and about 18K reads spread between 128 -
8191 usecs. Write traffic is largely clusstered in the 64 - 127 usecs bracket.
The faster read traffic is probably coming from a filesystem cache and the slower
traffic from disk. The reason why the writes are so consistenly fast is becuase
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

"because"


In this example you can see that the read traffic is rather bi-modal, with about
26K reads falling within 8 - 30 usecs and about 18K reads spread between 128 -
8191 usecs. Write traffic is largely clusstered in the 64 - 127 usecs bracket.
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

"clustered"

applications reading from a NFS share and can better expose problems
experienced by NFS clients.

Note that this only traces the common NFS operations (read,write,open and
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Spaces between operations: "read, write, open, and ..."

4 -> 7 : 137 |****************************************|
8 -> 15 : 3 | |

This shows a write workload, with writes hovering primarily in the 4-7ms range
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Period?

@samuelnair
Copy link
Contributor Author

Damn, I should have run it through ispell. Thanks for the spell check @goldshtn

@goldshtn
Copy link
Collaborator

LGTM! Thanks.

@goldshtn goldshtn merged commit ceff3e0 into iovisor:master Sep 12, 2017
CrackerCat pushed a commit to CrackerCat/bcc that referenced this pull request Jul 31, 2024
nfsdist: trace NFS operation latency distribution, similar to the other *dist tools.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants