Our DBAs have been doing some proof-of-concept work on Oracle on Red Hat linux
on System z compared to SUN.
They were running some tests which took a very long time on z compared to SUN.
I found out they were doing a lot of "grep" commands. I found that there were
many performance problems reported with grep when using utf-8 ( in general - not
limited to a particular platform). So, I got them to run the same test without
utf-8......
As per your findings about slowness with grep and utf8, I changed the LANG
variable on swplp18 (linux on z guest) from LANG=en_US.UTF-8 to LANG=C (for my
session)
This changes various locale settings.
The grep on swplp18 then ran in 6 seconds.
I set LANG back to utf8, and it ran in 24 seconds.
I ran the tests twice and it was about the same both times.
I tried the same thing on reg-data1 (linux on SUN x86), but either it's not a
factor there for some reason, or I'd have to have a much larger test to see the
difference
... ok, I tried a much larger test.
with LANG=C
18:31:11 to 18:40:41 9 min 30 sec
with LANG=en_US.UTF-8
18:42:21 to 18:52:16 9 min 55 sec
which is not that significant. I only did 1 test because I didn't want to wait
any more.
The version of grep is the same on both boxes.
Linux on z Guest:
Red Hat Enterprise Linux AS release 4 (Nahant Update 7) - z/VM 5.2 on a
z9BC
swplp18.swp.com:102>grep -V
grep (GNU grep) 2.5.1
Linux on SUN x86:
Red Hat Enterprise Linux WS release 4 on SUN x86
reg-oid1.swp.com:102>grep -V
grep (GNU grep) 2.5.1
Has anyone else seen performance issues with grep and utf-8? Suggestions?
This e-mail and any attachment(s) are confidential and may be privileged.
If you are not the intended recipient please notify me immediately by return
e-mail,
delete this e-mail and do not copy, use or disclose it.
----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
