First off I am just starting to learn DTrace. This came about as a way to diagnose low latency issues with a ZFS pool used to house LUNs for a VMware environment.
Anyways I grabbed the script iscsirwlat.d script from http://www.solarisinternals.com/wiki/index.php/DTrace_Topics_iSCSI As it prints out read and write latencies. The problem isn't with the script, my question is around the assumption that DTrace does not impact performance (or it is very negligible). When I run an I/O test against the ZPOOL from a VM guest I get a max read and write throughput (for a 1Gbit connection) and the latencies of 600 ms (as reported by ESX). I then run the test with the DTrace script mentioned above running and max throughput drops by around 20Mb/sec and latencies (as reported by ESX) jump 1000%. The only difference between this run and the first run is the DTrace script running. That leads me to believe the DTrace script is having a significant impact on performance. I am running Solaris 11 updated (just the other day) to the latest SRU 17.5.0. Thanks ------------------------------------------- dtrace-discuss Archives: https://www.listbox.com/member/archive/184261/=now RSS Feed: https://www.listbox.com/member/archive/rss/184261/25769126-e243886f Modify Your Subscription: https://www.listbox.com/member/?member_id=25769126&id_secret=25769126-8d47a7b2 Powered by Listbox: http://www.listbox.com
