Sun ZFS Storage 7420
 
Oracle ZFS Storage ZS3-2
 
Sun Storage 7110 Unified Storage System
 
Oracle ZFS Storage ZS4-4
 
Sun Storage 7210 Unified Storage System
 
Sun Storage 7410 Unified Storage System
 
Oracle ZFS Storage ZS3-4
 
Sun ZFS Storage 7120
 
Sun Storage 7310 Unified Storage System
 
Oracle ZFS Storage Appliance Racked System ZS4-4
 
Sun ZFS Storage 7320
 
Oracle ZFS Storage ZS3-BA
 

PLA-Support>Sun Systems>DISK>ZFS Storage>SN-DK: ZS
 

In this Document

Purpose

Details

References

Applies to:

Purpose

Oracle ZFS Storage Appliance provides a powerful analytics platform to monitor storage performance.

The Analytics platform comes with an extensive number of statistics that are easy to turn both on and off.

Individually the impact of any statistic is negligible, but collectively, these statistics can have a performance impact in a production environment and should be used judiciously.

This document describes best practices for Analytics usage to minimize the performance overhead.

Details

Analytics provides a comprehensive observability of system resources and protocols in the Storage Appliance.

It comes with a range of built-in statistics, from simple stats, such as, the total number of operations, to complex drilldowns, such as, the total number of operations on a file requested by a specific client.

The complex breakdowns are based on data collected via Dynamic Tracing (DTrace) by using 1000s of instrumented points (probes) placed at the System Level.

These probes have no cost when they are not enabled and make complex tracing accessible and easy to use.

If the number of breakdowns or type of data collected per second gets large, however, performance can be impacted.

The following section lists best practices to minimize performance overhead from analytics.

Best Practices:

1. Use latest Software: We are constantly making user and performance improvements to analytics. The Oracle ZFS Storage appliance firmware release version should be kept updated.

Please reference: Oracle ZFS Storage Appliance: Software Updates (Doc ID 2021771.1).

2. Set a default retention policy: Analytics retains all data to provide historical views of days, weeks and months. Limit the amount of data retained by keeping these to a reasonable level.

Please reference Sun Storage 7000 Unified Storage System: How to configure Analytics for dataset retention policies (Doc ID 1461959.1).

3. Turn on new statistics on “as needed” basis: It is good to keep production systems as simple as possible. The more metrics turned on, the more the system has to work.

Consider deleting datasets, which are no longer used. The default datasets, however, will not incur much overhead on the system.

4. Use the breakdowns selectively: Some breakdowns, such as, by file, by client, or latency can accumulate very fast.

They should only be enabled on a “as needed” basis for debugging performance and then suspended or destroyed after the issue has been investigated.

Datasets, such as, arc.accesses[file], nfs3.ops[client] are expensive to collect and can become very large.

5. Suspend Analytics during takeover/failback and upgrades: During critical operations, such as, takeover/failback and upgrade, analytics should be suspended.

6. Size of Statistics: Check incore and disk size of the datasets, available under Analytics>Datasets menu in the Appliance UI.

Consider deleting and recreating datasets if incore size exceeds 700 MB and disk size used gets large relative to total capacity, e.g. 2 GB for a single dataset.

References

This solution has no attachment