Appendix

Glossary

CPU busy percent formula

Time Adaptive Server spends executing tasks on the engine
Time the engine is available to Adaptive Server

CPU usage

Cumulative counts

Data cache efficiency formula

Hits per second for this data cache
MB in this data cache

Data cache hit percent formula

Hits
Hits + Misses

Device hit percent formula

I/O requests granted immediately
Total I/O requests

Lock status

• Held/Blocking -- the process owns the lock and at least one other process is waiting for that lock to be released.
• Held/Not Blocking -- the process owns the lock and no other processes are waiting for it.
• Requested/Blocked -- the process has requested a lock that another process currently holds.
• Requested/Not Blocked -- the process has requested a lock that is free.

Lock types

• Shared Page Lock - Adaptive Server applies this lock for read operations. If a shared lock has been applied to a data page, other transactions can also acquire a shared lock even when the first transaction is not finished. However, no transaction can acquire an exclusive lock on the page until all shared locks on it are released.
  By default, Adaptive Server releases shared page locks after the scan is complete on the page. It does not hold them until the statement completes or until the end of the transaction. Transactions that need an exclusive page lock wait or block for the release of the shared page locks before continuing.
• Shared Intent Table Lock - Adaptive Server applies a shared intent table lock on any table that has an open cursor or a shared page lock on one of its pages. Setting a shared intent table lock prevents other transactions from subsequently acquiring a shared table lock on the table. The intent lock is held as long as the concurrent page lock is in effect.
• Shared Table Lock -- similar to shared page lock except that it applies to the entire table.
• Shared Address Lock -- similar to a shared page lock, except that it applies to non-leaf index pages.
• Shared Row Lock -- similar to a shared page lock, except that it applies to a row in a table.
• Shared Next Key -- a shared lock on a key used during table insertions. This is an instant duration lock and is usually transparent to the user. However, it could be blocked by a range lock owned by another process doing a serializable read (isolation level 3).
• Exclusive Page Lock -- Adaptive Server applies these locks for data modification operations. When a transaction gets an exclusive lock, other transactions cannot acquire a lock of any kind on the page until the exclusive lock is released at the end of the transaction. Those other transactions wait or block until the exclusive lock is released.
• Exclusive Intent Table Lock -- Adaptive Server applies an exclusive intent table lock on any table that has an open cursor or an exclusive page lock on one of its pages. Setting an exclusive intent table lock prevents other transactions from subsequently acquiring an exclusive table lock on the table. An intent lock is held as long as the concurrent page lock is in effect.
• Exclusive Table Lock -- similar to an exclusive page lock, except this lock applies to an entire table.
• Exclusive Address Lock -- similar to an exclusive page lock, except that it applies to non-leaf index pages.
• Exclusive Row Lock -- similar to an exclusive page lock, except this lock applies to a row in a table.
• Update Page Lock -- Adaptive Server applies update locks during the initial portion of an update, delete, or fetch (for cursors declared for update) operation when the pages are being read. The update locks allow shared locks on the page, but do not allow other update or exclusive locks. This is an internal lock to help avoid deadlocks. Later, if the pages need to be changed and no other shared locks exist on the pages, Adaptive Server promotes the update locks to exclusive locks.
• Update Row Lock -- similar to an update page lock, except this lock applies to a row in a table.

Network memory formula

               default                      number
3      *    network packet       *          of user
                size                       connections

Prefetch formula

Number of reused pages referenced by Adaptive Server
Pages per buffer * Number of reused buffers

Procedure cache hit percent formula

Procedure execution requests - Requests requiring physical reads
Procedure execution requests

Process states

• Alarm Sleep -- waiting on an alarm.
• Background -- Adaptive Server process executing.
• Bad Status -- undetermined error condition.
• Infected -- tagged by Adaptive Server as unprocessable.
• Lock Sleep -- waiting on a lock acquisition.
• Recv Sleep -- waiting on a network read.
• Runnable -- waiting to run according to priority and availability of CPU.
• Running -- executing.
• Send Sleep -- waiting on a network send.
• Sleeping -- paused for any other reason not listed here, such as: waiting on device I/O (physical reads) or waiting for client activity.
• Stopped -- process terminated.
• Unknown -- process state undeterminable.
• Remote I/O -- process is waiting on a remote server to complete an operation.
• Sync Sleep -- waiting on a synchronization message from another process in the family.

Rates (number/second)

Count for the most recent sample
Number of seconds in a sample interval

Cumulative rates - average number of occurrences per second in the current session.

Cumulative count for the session
Number of seconds since monitor opened

Received packet size formula

Bytes received during this session
Packets received during this session

Sent packet size formula

Bytes sent during this session
Packets sent during this session

Server CPU percent formula

Time Adaptive Server spends executing tasks on all engines
Time engines were available