1The following are additional notes on the iotop program, 2 3 4* When using -P, how can a process exceed 100% I/O? 5 6These percentages are based on disk time. They are in terms of a single disk. 7 8200% could mean 2 disks @ 100%, or 4 @ 50%, or some such combination. 9 10I could have capped it at 100% by dividing by disk count. I didn't. Disk 11utilisation is an asymmetric resource (unlike CPUs, which are (mostly) 12symmetric), so it's unfair to divide by all the disks capacity as an 13application cannot use every disks capacity (eg, writing to a /opt disk only). 14 15Would it be wise to report utilisation as 10% of overall capacity, if it 16could mean that 1 disk was SATURATED out of ten? A value of 10% could 17understate the problem. 18 19Instead I add the utilisations and don't divide. 1 disk saturated out of 10 20would be reported as 100% utilisation. This has the danger of overstating 21the problem (consider all ten disks at 10% utilisation, this would also be 22reported as 100%). 23 24Nothing is perfect when you are summarising to a single value! 25 26 27 28* Beware of overcounting metadevices, such as SVM and Veritas. 29 30The current version of iotop reports on anything the kernel believes to be 31a block disk device. A problem happens when a metadevice contains physical 32disk devices, and iotop reports on activity to both the metadevice and 33the physical devices, which overcounts activity. 34 35Consider a metadevice that contains two physical disks which are both 36running at 100% utilised. iotop -P may report 300% utilisation, which is 37200% for the disks + 100% for the metadevice. We'd probably want to see 38a value of 200%, not 300%. Eliminating the counting of metadevices in DTrace 39isn't easy (without inelegant "hardwiring" of device types), however I do 40intend to find a way to fix this in future versions. 41 42 43