db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Bergquist, Brett" <BBergqu...@canoga.com>
Subject Re: Could use some thoughts on a problem with a query in production
Date Tue, 06 Jun 2017 17:53:00 GMT
Thanks Bryan.  I am looking at all things :)   The fact that it is not reproducible with a
copy of the database is a hinderance but also can indicate something external.

The interesting thing is that the same query with the same parameters was run 40 times over
about an 8 hour period with about 30 minutes apart.  Each of these hit the exactly the same
problem.   At the same time, the same query requesting different time periods worked perfectly.
  Insertion rates into the database also were not affected.

Does anyone know anything about the “cache cleaning”?   I have started to peak and poke
at what it does but I am wondering if there is something in there that might be triggering
the issue.   The reason that it is peaking my interest is that within the code where the BLOCKED
threads are occurring are checked for the “cleaning”.   These are the areas that the code
is waiting.   it is almost such that something gets lock stepped in which the code cannot
progress because of something holding a monitor and the monitor is is being released and acquired
almost once per result.   Also the fact that the CPU utilization of derby went from about
5% to about 45% during this time indicates something is being worked on or a busy wait being
performed.

On Jun 6, 2017, at 9:28 AM, Bryan Pendleton <bpendleton.derby@gmail.com<mailto:bpendleton.derby@gmail.com>>
wrote:

Hi Brett?

What's the underlying storage system like? In particular, is it an SSD
storage system? Is there any chance that you happened to hit a spot
in time where there was some non-Derby event which was causing the
storage system to misbehave?

I understand that certain SSD implementations, and certain file
systems, have to perform automatic maintenance activity on the
SSD media, such as "wear leveling", which can impose significant
overhead on the file system and cause normal use of the file system
to be VERY slow for a significant period.

Here's a relatively old article on the subject; you might try some web searching
to see if you can get more recent information.

http://www.computerworld.com/article/2524589/data-center/analysis--ssd-performance----is-a-slowdown-inevitable-.html

Anyway, I guess what I'm suggesting is that you consider investigating
whether a non-Derby external factor is at work on the machine which
is suffering the performance problems.

thanks,

bryan



________________________________
Canoga Perkins
20600 Prairie Street
Chatsworth, CA 91311
(818) 718-6300

This e-mail and any attached document(s) is confidential and is intended only for the review
of the party to whom it is addressed. If you have received this transmission in error, please
notify the sender immediately and discard the original message and any attachment(s).
Mime
View raw message