On Solaris 11, and using ZFS, we monitor the IOPS of the filesystems in different datasets using:

fsstat /apool/adataset/* 1 100

where apool is the ZFS pool containing a parent dataset adataset with several sub-datasets below. The command returns something like:

 new  name   name  attr  attr lookup rddir  read read  write write
 file remov  chng   get   set    ops   ops   ops bytes   ops bytes
7.33M 72.4K 7.30M 58.1M 29.3M   489M  428K 2.62M 20.6G 15.1M 2.45T /apool/adataset/num_1
1.33M 12.9K 2.52M 88.1M  9.7M   921M  228K 2.91M 30.2G  7.1M 1.25T /apool/adataset/num_2
  134     0   134   536   536  7.98K     0     0     0   134 17.6M /apool/adataset/num_1
  144     0   144   576   576  8.58K     0     0     0   144 18.9M /apool/adataset/num_2

in the above the first two lines are the total accumulated since system startup for both children datasets, and the two next ones are the current activity over the last 1s.

Could someone explain if the "lookup ops" activity impacts the read/write IOPS, or point me to some good reference?

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.