Community
- Command Central
- MVP Program
- Monthly Mission
- Blogs
- Groups
- Events
- Media Vault
Products
- Observability
- Network Management
- Application Management
- IT Security
- IT Service Management
- System Management
- Database Management
Content Exchange
- SolarWinds Platform
- Server & Application Monitor
- Database Performance Analyzer
- Server Configuration Monitor
- Network Performance Monitor
- Network Configuration Manager
- SQL Sentry
- Web Help Desk
Free Tools & Trials
Store

Mismatch in Disk Latency between DPA and Windows

Looking for some help on interpreting some disk latency numbers. DPA is reporting upwards of 800ms response. If I drill into the actual Storage IO charts, it shows up to 2.5s on one of my drives. Windows however is reporting 4ms response. Trying to understand the difference.

Find more posts tagged with

Accepted answers

mandevil

We query SQL Server itself for these numbers. So even if the OS is seeing low latency across all drives, SQL is apparently realizing high service/stall times when IO requests are being made. We query from the dm_io_virtual_file_stats DMV for these metrics. If you get to a period of time where you see we're showing high latency, I'd suggest going after the raw data (see table definition here sys.dm_io_virtual_file_stats (Transact-SQL) ).

Keep in mind this metric is a delta (so you will have to take multiple samples and do a little math on it).

So, now I'll go to my mantra - what are the waits telling you? 8 )

Also, have you drilled tighter into Storage I/O to see which files are experiencing the most latency?

Here's a video which may help somewhat explain what you are looking at: Utilize DPA Storage I/O to find Performance Issues - Videos | SolarWinds .

All comments

mandevil

Keep in mind this metric is a delta (so you will have to take multiple samples and do a little math on it).

So, now I'll go to my mantra - what are the waits telling you? 8 )

Also, have you drilled tighter into Storage I/O to see which files are experiencing the most latency?

Here's a video which may help somewhat explain what you are looking at: Utilize DPA Storage I/O to find Performance Issues - Videos | SolarWinds .

mprobus

Thank you for the information.

To answer your question, yes I was able to drill down to find the specific mdf having problems. It reported times of up to 4200ms. Most of the wait was pageiolatch.

Now that I understand where DPA gets it's metrics, I'm still not sure why there is such high latency in SQL when Windows doesn't have the problem. The storage team is not seeing latency on the backend either. Troubleshooting continues.

gcp

This article may give you some ideas on investigating pageiolatch: https://sqlperformance.com/2014/06/io-subsystem/knee-jerk-waits-pageiolatch-sh