Difference between revisions of "Information Systems:Analyzing disk usage on Bart"
m |
m |
||
| (6 intermediate revisions by the same user not shown) | |||
| Line 3: | Line 3: | ||
It should be a habit to reassess disk usage on Bart every once in a while. It helps to understand storage from a big-picture standpoint: large tables in our ERP, stale temporary files, IFS folders with accumulating log files etc. This article documents the different methods and tools that can help in such an analysis. According to IBM, these are some limits we should be aware of: |
It should be a habit to reassess disk usage on Bart every once in a while. It helps to understand storage from a big-picture standpoint: large tables in our ERP, stale temporary files, IFS folders with accumulating log files etc. This article documents the different methods and tools that can help in such an analysis. According to IBM, these are some limits we should be aware of: |
||
| + | [[File:bart-fs-limits.jpg|400px]] |
||
| + | Also, IBM also recommends that storage stays below the 70% usage mark. |
||
=Tools= |
=Tools= |
||
| + | ====WRKDSKSTS==== |
||
| + | Use this command to quickly see the disk usage in percentage. This also shows IO levels for each disk. |
||
====RTVDSKINF==== |
====RTVDSKINF==== |
||
| + | Retrieve Disk Info. [https://www.ibm.com/developerworks/ibmi/library/i-using-rtvdskinf-command/ This is a really good article] on RTVDSKINF and related tools used to diagnose disk usage health. This is submitted as a batch job and takes approximately 20-30 min. PRTDSKINF produces a report from this command. |
||
| − | Retrieve Disk Info. |
||
| − | |||
| + | ====SYSPARTITIONSTAT (SQL view)==== |
||
| + | This is another one of those IBM SQL Views that allow you to get system stats through SQL instead of green screen commands. This is useful for checking the largest files occupying a library, and the number of records in those files. |
||
====ANZLIBBRM==== |
====ANZLIBBRM==== |
||
Note: Run RTVDSKINF before running this tool! |
Note: Run RTVDSKINF before running this tool! |
||
| + | Analyze Libraries using BRMS. Uses info from RTVDSKINF to produce a report with useful info related to BRMS backups: estimated save times, estimated portion of the backup media occupied etc. |
||
=Observations and other notes= |
=Observations and other notes= |
||
| − | For the same BRMS job (or more accurately, for the same list of objects specified by the job to be saved), certain steps seem to take longer for no apparent reason. One possible theory however, is that on fresh media (monthly tapes), writing is faster than on recycled media (daily tapes). |
+ | * For the same BRMS job (or more accurately, for the same list of objects specified by the job to be saved), certain steps seem to take longer for no apparent reason. One possible theory however, is that on fresh media (monthly tapes), writing is faster than on recycled media (daily tapes). |
5/31/17 3:42:21 2 blocks processed for sequence 147, volume 000015, on de |
5/31/17 3:42:21 2 blocks processed for sequence 147, volume 000015, on de |
||
| Line 21: | Line 27: | ||
6/01/17 3:40:44 2 blocks processed for sequence 147, volume 001804, on de |
6/01/17 3:40:44 2 blocks processed for sequence 147, volume 001804, on de |
||
6/01/17 3:47:46 341 objects saved from library CVT454AP. |
6/01/17 3:47:46 341 objects saved from library CVT454AP. |
||
| + | |||
| + | * QUSRBRM is 54GB. Why? |
||
| + | [[Category: IBM i]] |
||
| + | [[Category: I.T. Periodic Tasks]] |
||
Latest revision as of 13:48, 1 June 2017
Overview
It should be a habit to reassess disk usage on Bart every once in a while. It helps to understand storage from a big-picture standpoint: large tables in our ERP, stale temporary files, IFS folders with accumulating log files etc. This article documents the different methods and tools that can help in such an analysis. According to IBM, these are some limits we should be aware of:
Also, IBM also recommends that storage stays below the 70% usage mark.
Tools
WRKDSKSTS
Use this command to quickly see the disk usage in percentage. This also shows IO levels for each disk.
RTVDSKINF
Retrieve Disk Info. This is a really good article on RTVDSKINF and related tools used to diagnose disk usage health. This is submitted as a batch job and takes approximately 20-30 min. PRTDSKINF produces a report from this command.
SYSPARTITIONSTAT (SQL view)
This is another one of those IBM SQL Views that allow you to get system stats through SQL instead of green screen commands. This is useful for checking the largest files occupying a library, and the number of records in those files.
ANZLIBBRM
Note: Run RTVDSKINF before running this tool!
Analyze Libraries using BRMS. Uses info from RTVDSKINF to produce a report with useful info related to BRMS backups: estimated save times, estimated portion of the backup media occupied etc.
Observations and other notes
- For the same BRMS job (or more accurately, for the same list of objects specified by the job to be saved), certain steps seem to take longer for no apparent reason. One possible theory however, is that on fresh media (monthly tapes), writing is faster than on recycled media (daily tapes).
5/31/17 3:42:21 2 blocks processed for sequence 147, volume 000015, on de 5/31/17 4:09:01 341 objects saved from library CVT454AP.
6/01/17 3:40:44 2 blocks processed for sequence 147, volume 001804, on de 6/01/17 3:47:46 341 objects saved from library CVT454AP.
- QUSRBRM is 54GB. Why?