When to Add the First Write Log Device (CLI)

Use this procedure to determine if you need a first write log device for Oracle ZFS Storage Appliance. Fibre Channel and iSCSI writes are synchronous and benefit from log devices when the write cache is enabled. Be sure to thoroughly read and understand the documented implications of enabling write caching on LUNs before proceeding. For more information, see Space Management for Shares in Oracle ZFS Storage Appliance Administration Guide, Release OS8.8.x.

To determine if you need to add more than one write log device, see When to Add More Write Log Devices (CLI).

1. Create a worksheet as described in Creating a Worksheet (CLI), select that worksheet, and then enter dataset.

   hostname:analytics worksheets> select worksheet-000
   hostname:analytics worksheet-000> dataset

2. Enter set name=nfs2.ops[op], and then enter commit to add NFSv2 operations per second broken down by type of operation to your worksheet.

   hostname:analytics worksheet-000 dataset (uncommitted)> set name=nfs2.ops[op]
                              name = nfs2.ops[op]
   hostname:analytics worksheet-000 dataset (uncommitted)> commit

3. Enter dataset.
4. Repeat steps 2 and 3 to add the following datasets:

   • NFSv3 operations per second, broken down by type of operation (nfs3.ops[op])

   • NFSv4 operations per second, broken down by type of operation (nfs4.ops[op])

   • NFSv4.1 operations per second, broken down by type of operation (nfs4-1.ops[op])

   • iSCSI operations per second, broken down by type of operation (iscsi.ops[op])

   • Fibre Channel operations per second, broken down by type of operation (fc.ops[op])

   • SMB operations per second, broken down by type of operation (smb.ops[op])

   hostname:analytics worksheet-000 dataset (uncommitted)> set name=nfs3.ops[op]
                              name = nfs3.ops[op]
   hostname:analytics worksheet-000 dataset (uncommitted)> commit
   hostname:analytics worksheet-000> dataset
   hostname:analytics worksheet-000 dataset (uncommitted)> set name=nfs4.ops[op]
                              name = nfs4.ops[op]
   hostname:analytics worksheet-000 dataset (uncommitted)> commit
   hostname:analytics worksheet-000> dataset
   hostname:analytics worksheet-000 dataset (uncommitted)> set name=nfs4-1.ops[op]
                              name = nfs4-1.ops[op]
   hostname:analytics worksheet-000 dataset (uncommitted)> commit
   hostname:analytics worksheet-000> dataset
   hostname:analytics worksheet-000 dataset (uncommitted)> set name=iscsi.ops[op]
                              name = iscsi.ops[op]
   hostname:analytics worksheet-000 dataset (uncommitted)> commit
   hostname:analytics worksheet-000> dataset
   hostname:analytics worksheet-000 dataset (uncommitted)> set name=fc.ops[op]
                              name = fc.ops[op]
   hostname:analytics worksheet-000 dataset (uncommitted)> commit
   hostname:analytics worksheet-000> dataset
   hostname:analytics worksheet-000 dataset (uncommitted)> set name=smb.ops[op]
                              name = smb.ops[op]
   hostname:analytics worksheet-000 dataset (uncommitted)> commit

5. Enter done, and then enter done again to exit the context.

   hostname:analytics worksheet-000> done
   hostname:analytics worksheets> done

6. Wait at least 15 minutes, and then go to analytics datasets.

   Note:

   Fifteen minutes is a general guideline. This amount of time may be adjusted if you have a performance-sensitive, short-duration synchronous write workload.

   hostname:> analytics datasets

7. Enter show to view a list of available datasets.

   hostname:analytics datasets> show
   Datasets:
   
   DATASET     STATE    INCORE  ONDISK  NAME
   dataset-000 active    1.27M   15.5M  arc.accesses[hit/miss]
   dataset-001 active     517K   9.21M  arc.accesses[hit/miss=metadata hits][L2ARC eligibility]
   ...
   dataset-030 active     290K   7.80M  nfs2.ops[op]
   hostname:analytics datasets>

8. Enter select and the dataset with the name nfs2.ops[op].

   In this example, dataset name nfs2.ops[op] corresponds to dataset-030.

   hostname:analytics datasets> select dataset-030

9. Enter read 900 to read the last 900 seconds, or 15 minutes, of the dataset, and copy and save the data to you environment for future reference.

   hostname:analytics dataset-030> read 900

10. Enter done.

    hostname:analytics dataset-030> done

11. Repeat steps 7 through 10 for the following datasets:

    • NFSv3 operations per second, broken down by type of operation (nfs3.ops[op])

    • NFSv4 operations per second, broken down by type of operation (nfs4.ops[op])

    • NFSv4.1 operations per second, broken down by type of operation (nfs4-1.ops[op])

    • iSCSI operations per second, broken down by type of operation (iscsi.ops[op])

    • Fibre Channel operations per second, broken down by type of operation (fc.ops[op])

    • SMB operations per second, broken down by type of operation (smb.ops[op])

    Note:

    Remember to copy and save the data for each dataset to your environment for future reference.

    hostname:analytics datasets> show
    ...
    hostname:analytics datasets> select dataset-032
    hostname:analytics dataset-032> read 900
    ...
    hostname:analytics dataset-032> done
    hostname:analytics datasets> show
    ...
    hostname:analytics datasets> select dataset-034
    ...
    hostname:analytics datasets> select dataset-27
    ...
    hostname:analytics datasets> select dataset-13
    ...
    hostname:analytics datasets> select dataset-07
    ...
    hostname:analytics datasets> select dataset-40

12. Examine the data.

    You might want to add the first write log device when one or all of the following conditions are present:

    • The sum of iSCSI writes, Fibre Channel writes, and NFS/SMB synchronous operations is at least 1000 per second

    • There are at least 100 NFS commits per second