QoS is not available in Data ONTAP 7-Mode systems; however, you can still collect some CLI data to review your current issue. Before you get into reviewing data, here are some common areas that might be beneficial:
- My CPU is high – is this bad?
For more information, see article 3014056: What does high CPU utilization indicate?
- Why are my HDDs busy?
For more information, see the HDD Latency Troubleshooting article – 1014701: How to assess disk-level response times
- Some users are seeing slowness accessing shares:
If none of these are applicable to your situation, call NetApp Support and open a support case. The following will help in case resolution:
- Detailed information regarding the issue at hand
- Shares, volumes, and nodes affected, and time stamps of when the issue occurred.
- Any error messages encountered, and where they were seen.
- Is the issue reproducible, random, or constant?
- When did the issue begin, and has anything in the environment changed recently?
- An example of a good problem statement:
“We are having an issue with increased read latency on volume XYZ at 9am this morning. Previously we saw 10ms, and now we see 40ms. This is impacting our main business application and causing us to lose revenue”
- If you are using a version of Data ONTAP earlier than 8.3, collect a Perfstat while the issue is occurring:
- If you cannot capture while the issue is resident, attempt to capture the workload during a point in time which most closely resembles the time during which the issue occurred
- How to collect a Perfstat? (-t5 and -i5 are suggested for the time and iteration variables) – For more information, see article 3014095: FAQ: Perfstat collection methods omnibus
- If you are at or above Data ONTAP 8.3, send in a PerfArchive that covers the timeframe in question.
- For more information on how to submit a PerfArchive, see article 3014366: FAQ: Performance Archives: Performance First Failure Data Collection for clustered Data ONTAP 8.3 and later