You are viewing a plain text version of this content. The canonical link for it is here.
Posted to server-dev@james.apache.org by Eugene Drobitko <eu...@luciddream.com> on 2003/06/03 18:52:04 UTC
Re[4]: Tests of mail list.
NJB> Please keep this discussion on the mailing list.
OK.
NJB> Version 2.2.0a3 has all of the changes.
NJB> Are you sure that you are running the new code?
Yes. The james-binary-2.2.0a3.tar.gz was downloaded and unpacked in
new separate directory from previous version .
NJB> If you just put the new SAR
NJB> file in place, it would not overwrite your existing sqlResources.xml file.
No. All files are new.
NJB> If you do not have a separate entry for:
NJB> <sql name="selectByLowercaseName" db="mysql">
NJB> SELECT username, pwdHash, pwdAlgorithm, useForwarding,
NJB> forwardDestination, useAlias, alias
NJB> FROM ${table}
NJB> WHERE username = ?
NJB> </sql>
The new used sqlResources.xml contains this.
NJB> then you have the old file. The normal selectByLowercaseName entry has:
NJB> WHERE lower(username) = ?
NJB> Using that WHERE clause disables the index. You should not only change the
NJB> create statement for McKoi, but also clone the selectByLowercaseName, using
NJB> whatever you setup as a dbmatcher, e.g.:
NJB> <sql name="selectByLowercaseName" db="mckoi">
NJB> SELECT username, pwdHash, pwdAlgorithm, useForwarding,
NJB> forwardDestination, useAlias, alias
NJB> FROM ${table}
NJB> WHERE username = ?
NJB> </sql>
Yes, this was made.
Also as modification of SQL strings "CREATE TABLE ... " for making
case-insensitive of "username" for McKoi only.
NJB> Also, simply changing the entries in sqlResources.xml will not change an
NJB> existing table. You have to modify it manually and rebuild the index, or
NJB> drop the table entirely and recreate it. In the case of MySQL, that isn't
NJB> necessary, but it is for McKoi.
All tables was dropped for McKoi and for MySQL also.
NJB> --- Noel
NJB> -----Original Message-----
NJB> From: Eugene Drobitko [mailto:eugene@luciddream.com]
NJB> Sent: Tuesday, June 03, 2003 11:43
NJB> To: Noel J. Bergman
NJB> Subject: Re[2]: Tests of mail list.
NJB>> Eugene,
NJB>> The CVS has changes, as does v2.2.0a2
NJB>> (http://jakarta.apache.org/builds/jakarta-james/latest/).
NJB> I found only the version james-2.2.0a3.
NJB> downloaded and tested.
NJB>> Give it a shot.
NJB>> You'll still see slow performance for getUserByName unless you
NJB> customize the
NJB>> sqlResources.xml file for McKoi.
NJB> Yes I saw no performance for McKoi.
NJB>> See http://mckoi.com/database/maillist/msg01981.html. When you have it
NJB> working
NJB>> well, please submit a patch to sqlResources for McKoi.
NJB> I changed sqlResources.xml for McKoi, as described.
NJB> as I understand, this changes for case insensitive in names.
NJB> Re-tested and but now no changed seen.
NJB> Then I tested this version for MySQL database.
NJB> For MySQL this version also provide the same results as (v2.2.0a1 or
NJB> early).
NJB> Maybe I got not that version? Where I can get a2?
NJB> Eugene
NJB> Best regards, Eugene mailto:eugene@luciddream.com
NJB> ---------------------------------------------------------------------
NJB> To unsubscribe, e-mail: james-dev-unsubscribe@jakarta.apache.org
NJB> For additional commands, e-mail: james-dev-help@jakarta.apache.org
Best regards, Eugene mailto:eugene@luciddream.com
---------------------------------------------------------------------
To unsubscribe, e-mail: james-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: james-dev-help@jakarta.apache.org
RE: Re[6]: Tests of mail list.
Posted by "Noel J. Bergman" <no...@devtech.com>.
Eugene,
Glad to hear that it is work well for you now. I used the default scripts,
but I do add:
PHOENIX_JVM_OPTS="-server"
to the environment when I run James.
--- Noel
-----Original Message-----
From: Eugene Drobitko [mailto:eugene@luciddream.com]
Sent: Tuesday, June 03, 2003 13:29
To: James Developers List
Subject: Re[6]: Tests of mail list.
Noel,
I rechecked all settings of both versions and find one distinction
It is settings for memory in run.bat for start James.
I set setting as previous version and see the increasing speed in 4-5
times.
Sorry .
___________________________________________________
v2.2.0a3 -Xms256M -Xmx1024M
peak memory usage - 282 432Kb
{21:09:14} 0 / - empty
{21:09:24} 0 / - empty
{21:09:34} 0 / - empty
{21:09:45} 0 / transport
{21:09:55} 695 / transport
{21:10:05} 1643 / transport
{21:10:15} 2500 / - empty
{21:10:25} 2500 / - empty
___________________________________________________
v2.2.0a3 default memory settings
peak memory usage - 87 304Kb
{21:14:33} 0 / - empty
{21:14:43} 4 / transport
{21:14:53} 182 / transport
{21:15:03} 398 / transport
{21:15:13} 610 / transport
{21:15:23} 825 / transport
{21:15:33} 1032 / transport
{21:15:43} 1236 / transport
{21:15:53} 1454 / transport
{21:16:03} 1679 / transport
{21:16:13} 1899 / transport
{21:16:23} 2123 / transport
{21:16:33} 2355 / transport
{21:16:43} 2500 / - empty
___________________________________________________
v2.2.0a1 -Xms256M -Xmx1024M
peak memory usage - 282548 Kb
{06:30:21} 0 / - empty
{06:30:31} 0 / transport
{06:30:41} 0 / transport
{06:30:51} 0 / transport
{06:31:01} 10 / transport
{06:31:11} 187 / transport
{06:31:21} 369 / transport
{06:31:31} 555 / transport
{06:31:41} 742 / transport
{06:31:51} 934 / transport
{06:32:01} 1127 / transport
{06:32:11} 1320 / transport
{06:32:21} 1494 / transport
{06:32:31} 1685 / transport
{06:32:41} 1877 / transport
{06:32:51} 2072 / transport
{06:33:01} 2267 / transport
{06:33:11} 2466 / transport
{06:33:21} 2500 / - empty
Eugene
---------------------------------------------------------------------
To unsubscribe, e-mail: james-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: james-dev-help@jakarta.apache.org
Re[6]: Tests of mail list.
Posted by Eugene Drobitko <eu...@luciddream.com>.
Noel,
I rechecked all settings of both versions and find one distinction
It is settings for memory in run.bat for start James.
I set setting as previous version and see the increasing speed in 4-5
times.
Sorry .
___________________________________________________
v2.2.0a3 -Xms256M -Xmx1024M
peak memory usage - 282 432Kb
{21:09:14} 0 / - empty
{21:09:24} 0 / - empty
{21:09:34} 0 / - empty
{21:09:45} 0 / transport
{21:09:55} 695 / transport
{21:10:05} 1643 / transport
{21:10:15} 2500 / - empty
{21:10:25} 2500 / - empty
___________________________________________________
v2.2.0a3 default memory settings
peak memory usage - 87 304Kb
{21:14:33} 0 / - empty
{21:14:43} 4 / transport
{21:14:53} 182 / transport
{21:15:03} 398 / transport
{21:15:13} 610 / transport
{21:15:23} 825 / transport
{21:15:33} 1032 / transport
{21:15:43} 1236 / transport
{21:15:53} 1454 / transport
{21:16:03} 1679 / transport
{21:16:13} 1899 / transport
{21:16:23} 2123 / transport
{21:16:33} 2355 / transport
{21:16:43} 2500 / - empty
___________________________________________________
v2.2.0a1 -Xms256M -Xmx1024M
peak memory usage - 282548 Kb
{06:30:21} 0 / - empty
{06:30:31} 0 / transport
{06:30:41} 0 / transport
{06:30:51} 0 / transport
{06:31:01} 10 / transport
{06:31:11} 187 / transport
{06:31:21} 369 / transport
{06:31:31} 555 / transport
{06:31:41} 742 / transport
{06:31:51} 934 / transport
{06:32:01} 1127 / transport
{06:32:11} 1320 / transport
{06:32:21} 1494 / transport
{06:32:31} 1685 / transport
{06:32:41} 1877 / transport
{06:32:51} 2072 / transport
{06:33:01} 2267 / transport
{06:33:11} 2466 / transport
{06:33:21} 2500 / - empty
Eugene
NJB> OK, sounds like you did all the necessary things that come to mind.
NJB> Please turn on DEBUG for the spool manager, and post a section related to
NJB> the processing of a single message through the spooler.
To make it all the same
Or it is not necessary yet?
NJB> Are you able to build James from source?
Yes. I downloaded today latest source when re-tested and compared
versions.
NJB> If necessary, I'll post to e-mail
NJB> a change to re-insert the timing code that I was using internally here to
NJB> measure sections of the process.
NJB> --- Noel
NJB> ---------------------------------------------------------------------
NJB> To unsubscribe, e-mail: james-dev-unsubscribe@jakarta.apache.org
NJB> For additional commands, e-mail: james-dev-help@jakarta.apache.org
Best regards, Eugene mailto:eugene@luciddream.com
---------------------------------------------------------------------
To unsubscribe, e-mail: james-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: james-dev-help@jakarta.apache.org
RE: Re[4]: Tests of mail list.
Posted by "Noel J. Bergman" <no...@devtech.com>.
OK, sounds like you did all the necessary things that come to mind.
Please turn on DEBUG for the spool manager, and post a section related to
the processing of a single message through the spooler.
Are you able to build James from source? If necessary, I'll post to e-mail
a change to re-insert the timing code that I was using internally here to
measure sections of the process.
--- Noel
---------------------------------------------------------------------
To unsubscribe, e-mail: james-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: james-dev-help@jakarta.apache.org