Troubleshooting
Problem
VMSTAT output can be used to check for Disk performance bottlenecks
Cause
Poor Disk performance
Diagnosing The Problem
VMSTAT command that can be used to help determine a disk bottleneck by looking at the block I/O:
vmstat -Itw 10
1) There are many parameters with this command that can be used but this is a good place to start.
2) This command has low overhead. This can run all the time and prune entries after N days.
Resolving The Problem
Look at the “b” column. (If you’re using “cooked filesystems”.) This indicates how many threads are blocked/bottlenecked on I/O.
Look at the "p" column if you are using raw Logical Volumes for the storage pools. This indicates how many threads are blocked/bottlenecked on I/O.
Note: Tivoli Storage Manager Version 6.x does not use raw Logical Volumes for the DB.
Sample VMSTAT output. Note Columns "Fault/CPU/TIME" have been snipped for brevity:
kthr memory page
----------- --------------------- ------------------------------------
r b p avm fre fi fo pi po fr sr
5 0 0 1989487 5587 4443 1774 0 5 1492 3684
5 1 0 1989487 5670 4217 1037 0 4 873 2002
2 2 0 1989487 5640 4493 913 4 2 806 1132
4 1 0 1989487 5665 4177 1428 1 7 1075 3309
4 1 0 1989487 5658 4130 924 4 15 678 4199
3 0 0 1989487 5639 4487 1002 0 5 777 3375
2 2 0 1989487 5585 4493 1852 1 6 1419 3979
4 1 0 1989487 5361 4444 1049 0 8 914 1291
3 1 0 1989487 5633 4335 1038 1 8 782 1023
Values of 0 are good. Look for consistent values over time that are greater than 0. This indicates the I/O threads that are blocked. Occasional spikes are normal and expected.
Product Synonym
TSM ITSM ADSM IBM Spectrum Protect
Was this topic helpful?
Document Information
Modified date:
17 June 2018
UID
swg21615034