Interpreting SAR Data
Tuning A Netra Webserver

Interpreting SAR Data

CPU Utilization
%usr - Portion of time the CPU is running is User Mode
%sys - Portion of time the CPU is running in System Mode
%wio - Portion of time the CPU is idle, with process waiting on block IO
%idle - Portion of time CPU is otherwise idle

Buffer Activity
bread/s - Reads per second of data between system buffers and disk or other block device

bwrit/s - Writes per second of data between system buffers and disk or other block device

lread/s - Total read accesses of system buffers
lwrit/s - Total write accesses of system buffers
%rcache - Read cache hit ratio
%wcache - Write cache hit ratio
pread/s - Reads per second from physical device
pwrit/s - Writes per second to physical device

System Call Activity
scall/s - system calls of all types
sread/s - System read calls per second
swrit/s - System write calls per second
fork/s - System fork calls per second
exec/s - System exec calls per second
rchar/s - Characters read by system calls (not including incoming and outgoing exec and fork calls)

wchar/s - Characters written by system calls (not including incoming and outgoing exec and fork calls)

Block Device Activity
Block device activity is reported for each block device. On our systems each partition is a block device, and this data may be difficult to decipher due to the use of meta devices in the DiskSuite mirroring

%busy - Portion of time the device was busy servicing a transfer request
avque - Average number of requests outstanding during the %busy time.
read/s - Number of read transfers from this device
write/s - Number of write transfers to this device
blks/s - Number of bytes transferred to or from this device in 512 byte
blocks avwait - Average wait time in milliseconds.
Avserv - Average service time in milliseconds.

Paging Activity
pgout/s - Page-out requests per second
ppgout/s - Pages paged-out per second
pgfree/s - Pages per second placed on the free list by the page stealing daemon.

pgscan/s - Pages per second scanned by the page stealing daemon. %ufs_ipf - The percentage of UFS inodes taken off the freelist by iget which had reusable pages associated with them. These pages are flushed and cannot be reclaimed by processes. Thus, this is the percentage of igets with page flushes.

atch/s - page faults per second that are satisfied by reclaiming a page currently in memory (attaches per second).

pgin/s - page-in requests per second.
ppgin/s - pages paged-in per second.
pflt/s - page faults from protection errors per second (illegal access to page) or "copy-on-writes".

vflt/s - address translation page faults per second (valid page not in memory).
slock/s - faults per second caused by software lock requests requiring physical I/O.

Swapping and Switching Activity
swpin/s - Number of swap in calls per second
swpot/s - Number of swap out calls per second
bswin/s - Amount of data swapped in per second in 512 byte blocks bswot/s - Amount of data swapped out per second in 512 byte blocks pswch/s - process switches (Context switches per second)
Kernel Memory Allocation Activity
Freemem - average pages available to user processes.
Freeswap - disk blocks available for page swapping.
sml_mem - Amount of memory (in bytes) KMA has for the memory pool for small requests.

Alloc(sml) - Amount of memory allocated to satisfy requests for small amounts of memory

Fail - Number of requests for small amounts of memory that failed (were not satisfied)

lg_mem - Amount of memory (in bytes) KMA has for the memory pool for large requests.

Alloc - Amount of memory allocated to satisfy requests for large amounts of memory

Fail - Number of requests for large amounts of memory that failed (were not satisfied)

ovsz_alloc - The amount of memory allocated for oversized requests.

Fail - The number of oversized requests which could not be satisfied (because oversized memory is allocated dynamically, there is no pool)

Message and Semaphore Activity
msg/s - Primitive messages per second
sema/s - Primitive Semaphores per second Queue Information
runq-sz - the average number of messages waiting in the CPU run queue %runocc - indicates how often the queue has something in it (run occupancy).

Table Information
Proc-sz - maximum process table size, along with the current number of processes on the system
inod-sz - The current size and maximum size of the inode table (inode cache).
file-sz - The current size and maximum size of the system file table
lock-sz - The current size and maximum size of the system lock table
TTY Device Activity Reported for each tty device
rawch/s - Character input rate
canch/s - Character input rate processed by canon
outch/s - Character output rate