
Migrating to Sybase Adaptive Server Enterprise 11.5 6-15
Adaptive Server Enterprise 11.5 If Your Current Version is 4.x, 10.x or 11.0.x
The following table shows you how to interpret these entries:
Keep a copy of the error log from the earlier server. After you
perform an upgrade on a test server, check the same three lines in the
11.5 error log. If you have added no memory at this point, the larger
server structures will have caused the amount of cache to decrease.
Convert the sizes of proc buffers, proc headers, and data cache
buffers in old and new error logs to megabytes and subtract the new
from the old. This is how much you need to increase memory to keep
the same amount of data cache.
2. Calculate the Growth of Objects in Procedure Cache
This method of comparing memory usage between old and new
servers can be used when you have upgraded a test or development
system.
In both versions of the server, run commands like the following:
1> use sybsystemprocs
2> go
1> select object_id(“sp_help”)
2> go
1> dbcc prodbuf(sybsystemprocs, <object_id>, 0
2> go
Table 6-5: Interpreting size of memory structures
Structure What it is Units
proc buffers
Contains information about a compiled
object. There is one proc buffer for each
instance of an object in memory. The
number of proc buffers is the maximum
number of procedures that can be in
memory at one time.
76 bytes each
proc headers
Contains the entire compiled object. A
large object may span more than one proc
header. Larger objects means that fewer
proc headers are available.
number of 2K pages
buffers in buffer
cache
Hold data pages read into memory. Pre-
11.0, all buffers are 2K.
number of 2K pages
default data
cache and named
caches
Post-11.0, multiple data caches may have
buffers of 2, 4, 8, and 16K (twice this size
on Stratus)
KB
Comentarios a estos manuales