Have you ever met resistance to using Oracle SQL Trace because the perceived performance penalty is too great?
Most of us have heard people express that fear.
Is it true? Well, not really. Find out more at the following article.
#oracle#pythian#tracing
https://t.co/CepQWew6oz
Some oracle F/S files were slow and erratic.
I needed something outside the database to show the issues. Less than a benchmark, but sufficient for showing the issue outside the db.
https://t.co/I1ksPC0OH5
@oraclebase Indeed. I raised on a couple years ago about a sqlplus issue. There is a bug, no action in some time.
And it is one that should be an easy fix.
@whiteboardcoder if there are ever spaces in the file names:
add -print0 to find
add -0 to xargs
sort by date, though this can be trickier than it appears
... | xargs ls -ldtar
sort by size, descending, get top 10
... | xargs ls -ldtS | head -10
@Python_Dv more aliases
o: load a set of directories into an array
g: goto to a directory in the array
add-time: alias for adding time to my journal
mrskew: oracle tool for profiling trace files.
@JonathanGennick That looks quite interesting. Usually my use of Python with outliers is to remove them, as outliers obliterate charts.
Some of these outliers are interesting and worth investigation, but time does not suggest permit
@TanelPoder I recall using blkparse, and feeling it was insufficient. The problem with tools like this for many folks is they are not used frequently enough to recall just what to look for.
If I have need of it again, I may look at developing something more user-friendly.