Reload Analyzer Manual

Tandem Reload Analyzer Manual Page 36 of 42
Longest chain 197 <--
Previously 15
Chain length histogram
1 0
2 - 10 0
11 -25 0
26 - 50 0
51 - 100 0
> 100 1
Avg blocks/chain 197 <--
Previously 1
% of total blks in avg chain < 100% <--
Previously 1%
\NONSTOP.$DATA.INSURNC.SALESDB completed at - 17:05.48
Reload Analyzer is stopping.
The values in the General Information section have not changed. In the Block
Information section, the number of index blocks and data blocks has decreased
significantly. The file now has 197 data blocks compared to 279 before and 10 index
blocks compared to 23 before. The file had three index levels before the reload and
now has two index levels.
Look at the Data Chain Information section. One chain is 197 data blocks long. An
organized file like this one will have one long chain. The average blocks per chain is
197 because the file has only one chain. The percentage of total blocks in the average
chain is 100 percent. This percentage is calculated by taking average blocks per chain
divided by the total data blocks. (The average blocks per chain is 197 divided by 197
total data blocks equals 100 percent.) A percentage of 100 signifies an organized file.
Before you reload a file, consider the following. The file in Example 2-6, SALESDB, is
a small file, and reloading it probably will not affect the system. This might not be true
with a large file. An online FUP RELOAD can take time, degrade OLTP performance,
and generate large quantities of audit. Before reloading a file, you should consider the
impact on the system. For more information on how to reload, see the Reload Analyzer
support note.
When to Reload
Reload Analyzer provides information to help you decide whether to reload a file, but it
cannot determine whether a file should be reloaded because there are no universally
accepted criteria for this decision. The criteria you need to weigh include the cost of
application performance degradation versus the cost of reloading and the technical
confidence level of the person reloading the file.
The most common reason for considering a reload is degradation of application
performance. For example, if performance has gradually worsened for an application
that sequentially reads a frequently updated file, reloading the file is likely to help. In
other cases, it might be less clear whether file disorganization is affecting your
application performance.