vwinfo Statistics Display
Here is an example of the statistics output from the vwinfo command:
+--------------------------------------+------------------------------------+
|stat |value |
+--------------------------------------+------------------------------------+
|memory.memcontext_allocated |30920304 |
|memory.memcontext_maximum |214748364 |
|memory.memcontext_peak |41797856 |
|memory.memcontext_virtual_allocated |30920304 |
|memory.memcontext_virtual_maximum |70368744177664 |
|memory.memcontext_virtual_peak |66593025904 |
|memory.index_allocated |3579844 |
|memory.update_allocated |0 |
|memory.update_maximum |53687091 |
|memory.committed_transactions |0 |
|memory.bufferpool_maximum |3148349440 |
|memory.other_allocated |19360 |
|bm.block_size |524288 |
|bm.group_size |8 |
|bm.columnspace_used_blocks |8 |
|bm.bufferpool_total_blocks |24197 |
|bm.bufferpool_free_blocks |24197 |
|bm.bufferpool_used_blocks |0 |
|bm.bufferpool_cached_blocks |0 |
|system.active_sessions |0 |
|system.log_file_size |33239111 |
|system.threshold_log_condense |33554432 |
|server.port_number |54375 |
|server.pid |12800 |
+--------------------------------------+------------------------------------+
Fields are as follows:
memory.memcontext_allocated
The total memory, in bytes, allocated and administered by Vector Server memory pooling. It includes the total memory used by currently running sessions and global server memory. It includes memory.updates_allocated, memory.committed_transactions, memory.index_allocated.
memory.memcontext_maximum
The maximum allowed amount of physical memory, in bytes. If this threshold is reached, some queries will stop working. It is configured with the [memory] max_memory_size configuration parameter.
memory.memcontext_peak
The maximum amount of memory, in bytes, allocated at one single moment by running sessions during the lifetime of the server.
This information is useful for dimensioning the system. If this amount is close to the maximum allowed amount of physical memory, the server may return out-of-memory (OOM) errors. Consider changing the maximum allowed amount of physical memory on the [memory]max_memory_size configuration parameter and possibly adding more physical memory to the server.
memory.memcontext_virtual_allocated
The amount of virtual memory, in bytes, allocated by the currently running queries. In most cases, this amount is higher than the amount of physical memory. It indicates the address space the server is using rather than an amount of allocated memory. This memory is allocated only when the server will use it, and is then displayed in memory.memcontext_allocated. The server allocates more address space than memory to avoid memory fragmentation and copying when certain memory areas grow. You can influence this behavior with the [memory]max_overalloc configuration parameter.
memory.memcontext_virtual_maximum
The maximum allowed amount of allocated virtual memory, in bytes
memory.memcontext_virtual_peak
The maximum amount of virtual memory, in bytes, allocated at one single moment by running sessions during the lifetime of the server. This value will always be lower than the maximum allowed amount of allocated virtual memory (memory.memcontext_virtual_maximum).
Note: If this amount is close to the maximum, the server is running out of address space. A possible solution is lowering or disabling overallocation through the [memory]max_overalloc configuration parameter.
memory.index_allocated
The total memory, in bytes, allocated for in-memory indexes (min-max indexes and secondary indexes, if supported)
memory.update_allocated
The total memory, in bytes, used by memory-buffered updates of the committed transactions. The amount of this memory can be reduced by the automatic
propagation mechanism (see
Management of In-memory Updates). It can also be freed manually by issuing the statement MODIFY...TO COMBINE; COMMIT; on tables with in-memory updates, which will
propagate the in-memory updates to disk (see
Propagate In-memory Updates to Disk).
memory.update_maximum
The maximum amount of memory, in bytes, that can be used by memory-buffered updates. If memory.update_allocated reaches this limit,
propagation (see
Management of In-memory Updates) is triggered. If propagation is not successful, and the limit is reached, the memory-buffered updates may stop being accepted.
memory.committed_transactions
The total memory, in bytes, used by stored committed transactions. Vector keeps some information about a transaction in memory as long as any overlapping transaction is still running. This information is needed for potential conflict resolution. When this value becomes too high, the user must commit or abort some or all current sessions to free the memory occupied by the committed transactions. A typical cause for this value becoming high is a single old transaction that is still active.
memory.bufferpool_maximum
The total size, in bytes, of the buffer pool (disk block cache)
memory.other_allocated
(Linux only) Miscellaneous memory allocated, in bytes, outside the memory pooling.
bm.block_size
The disk block size used by this database. It is configured with the [cbm] block_size configuration parameter.
Note: Changing this parameter is possible only before creating a database. Later changes are ignored.
bm.group_size
The group size used by this database. It is configured with the [cbm] group_size configuration parameter.
Note: Changing this parameter is possible only before creating a database. Later changes are ignored.
bm.columnspace_used_blocks
The number of blocks used in total over all files and all locations
bm.bufferpool_total_blocks
The size of a buffer pool (disk data cache), in blocks. You can compute the total allowed physical size by multiplying this value by bm.block_size. This value is a sum of bm.bufferpool_free_blocks and bm.bufferpool_used_blocks.
bm.bufferpool_free_blocks
The number of blocks in the buffer pool that can be used for new data
bm.bufferpool_used_blocks
The number of blocks in the buffer pool that are currently locked by queries. This includes only blocks that are currently being used by scans and write to disk operations, which typically operate on only one block per column at a time. For statistics on cached data, see bm.bufferpool_cached_blocks.
bm.bufferpool_cached_blocks
The number of blocks in the buffer pool that contain cached data from disk
system.active_sessions
The number of currently running queries
system.log_file_size
Current size, in bytes, of the transaction log file
system.threshold_log_condense
Size, in bytes of the transaction log file at which the system tries to compact it
server.port_number
The port used for internal communication between the client and server
server.pid
The system process ID of the x100_server (iix100) process