- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
statistics show-periodic -object volume -instance volume_name -counter nfs_write_latency
I am running this command on volumes, and got different values, from 0 us to a few hundreds,
How high it is too high, Anybody please let me know your measurement? Thanks!
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi,
How high is too high: this depends on the application hosted on the particular nfs volume, and how much load is there from other volume on the controller.
In your case, I see sub milli second write latency which is fairly good for NFS volumes.Values lesser than 5 ms is ideally good.
Thanks
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Less than 5 ms (millisecond) on a VOLUME is good? are you sure?
we have 0 - hundreds us (microsecond), the maximum one is 1400 us, but the performace is terrible.
How to explain this?
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi,
0 to hundreds of micro seconds (Less than 1 ms) is ideally good, but again, depends on your workload.
I would stongly recommend you to open a support ticket to address this issue related to performance.
Thanks