Activelogsize space allocation

Answer

As a best practice the active log filesystem should have additional space (for failovers - see note below) but is not a requirement. Per the Tivoli Storage Manager Installation guide the Filesystem can be the same size, larger but not smaller than the space specified by the activelogsize.

Active and archive log space
When estimating space requirements for active and archive logs, include some
extra space for contingencies such as occasional heavy workloads and failovers.
...
v Ensure that the directory that holds the active log is as large as, or larger than,
the size of the active log. A directory that is larger than the active log can
accommodate failovers, if they occur.

However, having the activelogsize set to the same size as the space available on the Filesystem can still get an out of space condition (enospc). This occurs due to OS overhead (seen on Linux) for each log file in the active log directory. This may or may not also apply to other OS's.

Example:
The file system has 50GB of space. The activelogsize is set to 50GB. This in theory will create 100 - 512MB active log files. The problem is the OS uses 8k (8196 bytes) as overhead for each log file. So instead of the logs being 512MB they will be 512MB + 8k. This will leave only enough space to create 99 active log files. The server/DB2 will expect sufficient space for the 100 log files which can results in an out of space condition.


1024 x 1024 x 512 = 536870912 (512MB)
536879104 - 536870912 = 8192 bytes of overhead per log

Note on "failover":
Failover in this sense is in the event the ARCHIVElogspace fills and there is no failoverarchive log specified, DB2 tries to "save itself." It will actually start allocating and formatting new active logs beyond the activelogsize (in the active log directory) and attempt to clear the archive logs by doing a DB backup. As a best practice it is recommended to have the active log File System larger (by 20%) than what is specified by the activelogsize server option.