mturk 2005/09/23 00:32:24
Modified:jk/native/common jk_lb_worker.c
Log:
Fix compile time warnings. No functional change.
Revision ChangesPath
1.96 +4 -4 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
Hi, we have received
your request and will contact you as soon as possible.
Thank you.
The Dreamsailing Staff
www.dreamsailing.it
[EMAIL PROTECTED]
---
DreamSailing s.r.l.
Gruppo Metauro Mare
- Milano
Corso di Porta Romana, 101 - 20122 Milano
Tel. +39 02
Hi, we have received
your request and will contact you as soon as possible.
Thank you.
The Dreamsailing Staff
www.dreamsailing.it
[EMAIL PROTECTED]
---
DreamSailing s.r.l.
Gruppo Metauro Mare
- Milano
Corso di Porta Romana, 101 - 20122 Milano
Tel. +39 02
Hi, we have received
your request and will contact you as soon as possible.
Thank you.
The Dreamsailing Staff
www.dreamsailing.it
[EMAIL PROTECTED]
---
DreamSailing s.r.l.
Gruppo Metauro Mare
- Milano
Corso di Porta Romana, 101 - 20122 Milano
Tel. +39 02
Hi, we have received
your request and will contact you as soon as possible.
Thank you.
The Dreamsailing Staff
www.dreamsailing.it
[EMAIL PROTECTED]
---
DreamSailing s.r.l.
Gruppo Metauro Mare
- Milano
Corso di Porta Romana, 101 - 20122 Milano
Tel. +39 02
Hi, we have received
your request and will contact you as soon as possible.
Thank you.
The Dreamsailing Staff
www.dreamsailing.it
[EMAIL PROTECTED]
---
DreamSailing s.r.l.
Gruppo Metauro Mare
- Milano
Corso di Porta Romana, 101 - 20122 Milano
Tel. +39 02
Hi, we have received
your request and will contact you as soon as possible.
Thank you.
The Dreamsailing Staff
www.dreamsailing.it
[EMAIL PROTECTED]
---
DreamSailing s.r.l.
Gruppo Metauro Mare
- Milano
Corso di Porta Romana, 101 - 20122 Milano
Tel. +39 02
Hi, we have received
your request and will contact you as soon as possible.
Thank you.
The Dreamsailing Staff
www.dreamsailing.it
[EMAIL PROTECTED]
---
DreamSailing s.r.l.
Gruppo Metauro Mare
- Milano
Corso di Porta Romana, 101 - 20122 Milano
Tel. +39 02
Hi, we have received
your request and will contact you as soon as possible.
Thank you.
The Dreamsailing Staff
www.dreamsailing.it
[EMAIL PROTECTED]
---
DreamSailing s.r.l.
Gruppo Metauro Mare
- Milano
Corso di Porta Romana, 101 - 20122 Milano
Tel. +39 02
Hi, we have received
your request and will contact you as soon as possible.
Thank you.
The Dreamsailing Staff
www.dreamsailing.it
[EMAIL PROTECTED]
---
DreamSailing s.r.l.
Gruppo Metauro Mare
- Milano
Corso di Porta Romana, 101 - 20122 Milano
Tel. +39 02
Hi, we have received
your request and will contact you as soon as possible.
Thank you.
The Dreamsailing Staff
www.dreamsailing.it
[EMAIL PROTECTED]
---
DreamSailing s.r.l.
Gruppo Metauro Mare
- Milano
Corso di Porta Romana, 101 - 20122 Milano
Tel. +39 02
mturk 2005/09/23 00:18:21
Modified:jk/native/common jk_lb_worker.c
Log:
Fix loadbalancer retries to allow usage of cachesize lower then
number of threads. In case the worker is busy having no free endpoints,
and the retries config value is larger the total number of workers in
mturk 2005/09/15 22:52:26
Modified:jk/native/common jk_lb_worker.c jk_shm.c jk_shm.h
jk_status.c
Log:
Fix bug #36525 by properly aligning shared memory header struct.
This patch was provided and investigated by Rainer Jung. Thaks!
Also bring back 64 b
mturk 2005/09/12 05:53:06
Modified:jk/native/common jk_lb_worker.c jk_shm.c jk_shm.h
jk_status.c
Log:
Fix bug #36525 by not using 64 bit ints in shared memory.
For some strange reasons both Solaris and Irix are core dumping.
To deal with long running dat
Henri Gomez wrote:
ok, but couldn't build now from CVS, but it should works after
jk_u64_t is defined as unsigned long long.
Since we couldn't use portable.h on iSeries, it should be elsewhere ...
Sure... I will add the need #if defined(AS400).
2005/6/14, jean-frederic clere <[EMAIL PROTE
ok, but couldn't build now from CVS, but it should works after
jk_u64_t is defined as unsigned long long.
Since we couldn't use portable.h on iSeries, it should be elsewhere ...
>2005/6/14, jean-frederic clere <[EMAIL PROTECTED]>:
> Henri Gomez wrote:
> > status_strfsize ? APR ?
>
> Nearly: jaka
unsigned long is 32 bits and unsigned long long is 64 bits on iSeries
2005/6/14, Henri Gomez <[EMAIL PROTECTED]>:
> status_strfsize ? APR ?
>
> 2005/6/14, jean-frederic clere <[EMAIL PROTECTED]>:
> > Henri Gomez wrote:
> > > What do you need on iSeries ?
> >
> > Just to know what to use to have:
Henri Gomez wrote:
status_strfsize ? APR ?
Nearly: jakarta-tomcat-connectors/jk/native/common/jk_status.c ;-)
2005/6/14, jean-frederic clere <[EMAIL PROTECTED]>:
Henri Gomez wrote:
What do you need on iSeries ?
Just to know what to use to have:
JK_UINT4 (unsigned long of 32 bits) and
status_strfsize ? APR ?
2005/6/14, jean-frederic clere <[EMAIL PROTECTED]>:
> Henri Gomez wrote:
> > What do you need on iSeries ?
>
> Just to know what to use to have:
> JK_UINT4 (unsigned long of 32 bits) and JK_UINT8 (unsigned long long of 64
> bits)
> and to check status_strfsize().
>
> >
Henri Gomez wrote:
What do you need on iSeries ?
Just to know what to use to have:
JK_UINT4 (unsigned long of 32 bits) and JK_UINT8 (unsigned long long of 64 bits)
and to check status_strfsize().
2005/6/14, jean-frederic clere <[EMAIL PROTECTED]>:
Mladen Turk wrote:
jean-frederic cler
What do you need on iSeries ?
2005/6/14, jean-frederic clere <[EMAIL PROTECTED]>:
> Mladen Turk wrote:
> > jean-frederic clere wrote:
> >
> >> Mladen Turk wrote:
> >>
> >>>
> >>> Yes, but all that we need is 32 bit unsigned integer for JK_UINT4
> >>> What will you use if the int is 64 bits?
> >>
>
Mladen Turk wrote:
jean-frederic clere wrote:
Mladen Turk wrote:
Yes, but all that we need is 32 bit unsigned integer for JK_UINT4
What will you use if the int is 64 bits?
a long ;-)
Right :)
You will need the portable.in in that case, right?
Yes and I will ask Henri to check AS400
jean-frederic clere wrote:
Mladen Turk wrote:
Yes, but all that we need is 32 bit unsigned integer for JK_UINT4
What will you use if the int is 64 bits?
a long ;-)
Right :)
You will need the portable.in in that case, right?
---
Mladen Turk wrote:
jean-frederic clere wrote:
Mladen Turk wrote:
Perhaps u_int32_t would be more portable.
I would prefer to add in configure something like:
+++
AC_CHECK_SIZEOF(char, 1)
AC_CHECK_SIZEOF(int, 4)
AC_CHECK_SIZEOF(long, 4)
AC_CHECK_SIZEOF(short, 2)
AC_CHECK_SIZEOF(long long,
jean-frederic clere wrote:
Mladen Turk wrote:
Perhaps u_int32_t would be more portable.
I would prefer to add in configure something like:
+++
AC_CHECK_SIZEOF(char, 1)
AC_CHECK_SIZEOF(int, 4)
AC_CHECK_SIZEOF(long, 4)
AC_CHECK_SIZEOF(short, 2)
AC_CHECK_SIZEOF(long long, 8)
+++
and testing $a
Mladen Turk wrote:
jean-frederic clere wrote:
William A. Rowe, Jr. wrote:
Mladen; are you sure you weren't looking for 'long long', al la
int64_t? Falling over to the FPU is rarely the best
performance decision.
A little more coding is needed, because I have a related error (on
Reliant
jean-frederic clere wrote:
William A. Rowe, Jr. wrote:
Mladen; are you sure you weren't looking for 'long long', al la
int64_t? Falling over to the FPU is rarely the best
performance decision.
A little more coding is needed, because I have a related error (on
ReliantUnix):
undefined
William A. Rowe, Jr. wrote:
Mladen; are you sure you weren't looking for 'long long',
al la int64_t? Falling over to the FPU is rarely the best
performance decision.
A little more coding is needed, because I have a related error (on ReliantUnix):
+++
/home/apache20/apache20/build/libtool --si
Mladen; are you sure you weren't looking for 'long long',
al la int64_t? Falling over to the FPU is rarely the best
performance decision.
Bill
At 02:55 AM 6/13/2005, [EMAIL PROTECTED] wrote:
>mturk 2005/06/13 00:55:51
>
> Modified:jk/native/common jk_lb_worker.c jk_shm.h jk_status.c
Mladen,
I appreciate you addressing this issue. But isn't this fix just
delaying the problem? An IEEE double has 52 bits of mantissa versus
32 bits for size_t. So the problem will now occur at 4 PB
(petabytes). Sure, that's a lot of bytes and my webservers would
never see this kin
mturk 2005/06/13 00:55:51
Modified:jk/native/common jk_lb_worker.c jk_shm.h jk_status.c
Log:
Use double instead size_t for trensferred/read, so that lb doesn't
break on trnasferred mode when 2G of data has been send/read.
Revision ChangesPath
1.90 +5 -5 jaka
mturk 2005/05/15 09:33:47
Modified:jk/native/common jk_lb_worker.c
Log:
Do not log return value from endpoint. No functional change.
Revision ChangesPath
1.89 +3 -3 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
mturk 2005/05/15 08:22:06
Modified:jk/native/common jk_lb_worker.c jk_lb_worker.h jk_status.c
jk_util.c jk_util.h
Log:
Added lock directive for load balancer that allows more acurate
load balancing in case of burst load. Also display both method and
lock
mturk 2005/05/15 06:02:39
Modified:jk/native/common jk_lb_worker.c jk_worker.c jk_worker.h
Log:
Add use_map so that load balancer members do not inherit
mount directives if specified unless they are in worker.list.
Revision ChangesPath
1.87 +2 -2 jakarta-tom
mturk 2005/05/15 05:47:24
Modified:jk/native/common jk_lb_worker.c
Log:
Log error before closing workers.
Revision ChangesPath
1.86 +2 -3 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
===
mturk 2005/05/15 05:45:50
Modified:jk/native/common jk_lb_worker.c
Log:
Log worker creation as error instead as debug message.
Revision ChangesPath
1.85 +4 -5 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
==
mturk 2005/05/15 04:23:50
Modified:jk/native/common jk_lb_worker.c jk_shm.h jk_status.c
Log:
Add max_busy count to give the user an option to see what was
the maximum number of concurrent connections to Tomcat.
Revision ChangesPath
1.84 +5 -1 jakarta-tomcat-
mturk 2005/05/10 02:23:33
Modified:jk/native/common jk_lb_worker.c
Log:
Fix bug 34833 making failover to work.
Revision ChangesPath
1.83 +2 -2 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
==
pero2005/05/05 12:48:03
Modified:jk/native/common jk_lb_worker.c
Log:
Fix the sepcial only one worker case with stopped flag.
When disabled is true also new session can be created!!
Revision ChangesPath
1.82 +2 -2 jakarta-tomcat-connectors/jk/native/common
pero2005/05/05 08:40:17
Modified:jk/native/common jk_lb_worker.c
Log:
Don't use work when is stopped!
Revision ChangesPath
1.81 +2 -2 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
=
mturk 2005/05/05 05:48:35
Modified:jk/native/common jk_lb_worker.c
Log:
Allow setting 'secret' inside load balancer worker that will
set the secret for all its members if the membes have no specified
secret property.
Revision ChangesPath
1.80 +11 -1 jakarta
pero2005/04/26 08:28:18
Modified:jk/native/common jk_lb_worker.c jk_shm.h jk_status.c
jk_uri_worker_map.h jk_util.c jk_util.h
Log:
Add stopped flag for better cluster support to worker.
Many thanks to Mladen :->
Revision ChangesPath
1.79
mturk 2005/04/13 05:40:03
Modified:jk/native/common jk_lb_worker.c
Log:
Add difftime instead simply calculating time.
Revision ChangesPath
1.76 +2 -2 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
===
glenn 2005/04/09 08:29:14
Modified:jk/native/common jk_lb_worker.c
Log:
BUG: 34358 loadbalance.method=[T]raffic not enbaled
The lbworker->lbmethod was not being intialized from the
worker.properties config.
Fixed bug based on mladen's recommendation.
Revision Chan
mturk 2005/03/26 00:44:11
Modified:jk/native/common jk_lb_worker.c
Log:
In case ajp13 worker returns client error do not mark the worker
in error state. This operation is not recoverable, so return BAD_REQUEST
to web server.
Revision ChangesPath
1.73 +50 -24
mturk 2005/02/25 23:16:54
Modified:jk/native/common jk_lb_worker.c
Log:
Add failover for hot standby. This allows new sessions to use failover
worker or worker group. Limit is that you can not have two different
failover workers or workers groups.
Revision ChangesPath
Return Receipt
Your Re: cvs commit: jakarta-tomcat-connectors/jk/native/common
document jk_lb_worker.c jk_util.c jk_util.h
:
mturk 2005/02/22 00:59:57
Modified:jk/native/common jk_lb_worker.c
Log:
Update code style to be consistent. No functional change.
Revision ChangesPath
1.71 +3 -6 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
===
mturk 2005/02/22 00:42:45
Modified:jk/native/common jk_lb_worker.c jk_lb_worker.h jk_util.c
jk_util.h
Log:
Move "recover_time" to jk_util.c to be consistent with how other
properties are handled. No functional change.
Revision ChangesPath
1.70
mturk 2005/02/22 00:30:46
Modified:jk/native/common jk_lb_worker.c
Log:
Remove unused code. No functional change.
Revision ChangesPath
1.69 +1 -6 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
===
mturk 2005/02/22 00:25:29
Modified:jk/native/common jk_lb_worker.c
Log:
Set found worker to NULL when iterating over multiple session id's.
Revision ChangesPath
1.68 +2 -1 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
=
mturk 2005/02/21 23:26:55
Modified:jk/native/common jk_lb_worker.c
Log:
Fix infinite loop when all workers are in error state and sticky_session_force
is defined. If we found a worker by session id, and its in error state, return
no worker if sticky_session_force is defined (fo
mturk 2005/02/21 09:10:00
Modified:jk/native/common jk_lb_worker.c
Log:
Revert the latest patch. The factor is already calculated.
What was missnig was the lbmethod check.
Revision ChangesPath
1.66 +3 -18 jakarta-tomcat-connectors/jk/native/common/jk_lb_worke
mturk 2005/02/21 09:02:50
Modified:jk/native/common jk_lb_worker.c
Log:
Update lb_values for session workers.
Revision ChangesPath
1.65 +17 -1 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
===
mturk 2005/02/21 08:14:13
Modified:jk/native/common jk_lb_worker.c
Log:
Fix missing service jvm_route settings.
Revision ChangesPath
1.64 +5 -3 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
=
mturk 2005/02/19 00:12:30
Modified:jk/native/common jk_lb_worker.c jk_shm.h jk_status.c
Log:
Add worker is_busy runtime status flag. When we can not obtain a free
endpoint, mark the worker as busy rather then in error, so that recovery
is not needed.
Revision ChangesP
this message contains virus. don't send messages!
- Original Message -
From: <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Wednesday, February 16, 2005 2:55 PM
Subject: cvs commit: jakarta-tomcat-connectors/jk/native/common
jk_lb_worker.c jk_util.c jk_util.h
>
mturk 2005/02/17 05:48:47
Modified:jk/native/common jk_lb_worker.c
Log:
Update number of used endpoinds for each elected worker.
Revision ChangesPath
1.62 +3 -1 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
mturk 2005/02/17 05:41:04
Modified:jk/native/common jk_lb_worker.c jk_status.c
Log:
Display the number of workers currently serving requests.
Revision ChangesPath
1.61 +6 -2 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
mturk 2005/02/17 04:29:00
Modified:jk/native/common jk_lb_worker.c
Log:
Change some log messages. No functional change.
Revision ChangesPath
1.60 +5 -2 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
=
mturk 2005/02/17 01:12:24
Modified:jk/native/common jk_lb_worker.c jk_status.c
Log:
Remove unused variables. No functional change.
Revision ChangesPath
1.59 +1 -3 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
==
mturk 2005/02/17 00:28:41
Modified:jk/native/common jk_lb_worker.c
Log:
Small performance tune for balancers having single worker.
There is no need to find best worker if there is only one.
Single worker is used for workers that need full management.
Revision ChangesP
mturk 2005/02/16 23:16:32
Modified:jk/native/common jk_lb_worker.c jk_lb_worker.h
Log:
Set service route to either worker name or worker domain, depending
on how the worker has been elected.
Revision ChangesPath
1.57 +27 -19jakarta-tomcat-connectors/jk/native
mturk 2005/02/16 05:46:30
Modified:jk/native/common jk_lb_worker.c
Log:
If we can't get the endpoint try another worker by checking is_recoverable
only when we got the endpoint and endpoint service has been called.
Revision ChangesPath
1.55 +3 -3 jakarta-tom
mturk 2005/02/16 04:00:19
Modified:jk/native/common jk_lb_worker.c
Log:
Add traffic method for sticky sessions too.
Revision ChangesPath
1.54 +20 -6 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
=
mturk 2005/02/16 01:25:35
Modified:jk/native/common jk_lb_worker.c jk_util.c jk_util.h
Log:
Added disabled boolean directive to worker. This is used for
hot-standby workers that can be later enabled using jkstatus console.
Revision ChangesPath
1.53 +3 -1 jak
mturk 2005/02/16 00:30:58
Modified:jk/native/common jk_lb_worker.c
Log:
You already know ... Remove the CRLFs. SVN, you are my last chance!
Revision ChangesPath
1.52 +130 -130 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
=
mturk 2005/02/16 00:27:53
Modified:jk/native/common jk_lb_worker.c jk_lb_worker.h
Log:
Rewrite load balancer. Added byrequest and bytraffic methods.
Also keep session and domain models. Added sticky_session_force,
for non replicated Tomcats.
Revision ChangesPath
1.5
mturk 2005/02/14 23:25:41
Modified:jk/native/common jk_lb_worker.c
Log:
Add redirect rule to load balancer. Like domain if redirect is set and there
is sticky session it will be redirected to that worker. This will work only if
the worker is disabled, meaning not accepting new
mturk 2005/02/14 09:22:36
Modified:jk/native/common jk_lb_worker.c jk_lb_worker.h
Log:
Make lb_worker_t structure public so jkstatus can us it.
Revision ChangesPath
1.48 +7 -30 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
mturk 2005/02/13 04:21:57
Modified:jk/native/common jk_lb_worker.c
Log:
Use shm_alloc_worker, so that worker get it's id set, and that we can
monitor the total number of workers.
Revision ChangesPath
1.47 +3 -3 jakarta-tomcat-connectors/jk/native/common/jk_l
mturk 2005/02/12 11:22:35
Modified:jk/native/common jk_lb_worker.c
Log:
Log as error if shm_alloc fails.
Revision ChangesPath
1.46 +3 -2 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
mturk 2005/02/12 08:59:45
Modified:jk/native/common jk_lb_worker.c
Log:
Check if malloc has failed before using that memory, not after.
Revision ChangesPath
1.45 +6 -5 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
=
mturk 2005/02/12 03:32:13
Modified:jk/native/common jk_lb_worker.c jk_shm.c jk_shm.h
Log:
Simplify shared memory. If shmem is not defined use plain jk_pool instead.
Revision ChangesPath
1.44 +5 -5 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
mturk 2005/02/09 04:34:13
Modified:jk/native/common jk_lb_worker.c
Log:
Use shared memory for lb_worker shared data too. This will allow to
change sticky session processing at runtime.
Revision ChangesPath
1.43 +33 -29jakarta-tomcat-connectors/jk/native/commo
mturk 2005/02/09 03:46:57
Modified:jk/native/common jk_lb_worker.c
Log:
Use shared memory for worker record data. Now the runtime info
is shared among all child processes.
Revision ChangesPath
1.42 +68 -74jakarta-tomcat-connectors/jk/native/common/jk_lb_worke
mturk 2004/12/17 00:33:43
Modified:jk/native/common jk_lb_worker.c
Log:
Search for domain types 3 and 4 only if the worker has local workers
defined. They'll return null worker if not defined in any case.
Revision ChangesPath
1.38 +22 -18jakarta-tomcat-connec
mturk 2004/12/16 07:07:55
Modified:jk/native/common jk_lb_worker.c
Log:
Reorganize logging for getting candidate worker. It is not an error
if the domain is for the worker is not set.
Revision ChangesPath
1.37 +5 -8 jakarta-tomcat-connectors/jk/native/common
mturk 2004/12/16 02:56:03
Modified:jk/native/common jk_lb_worker.c
Log:
Make load balancer thread safe, since we can have a cross request
corruption of data, so that we don't use the worker that some other
thread marked as in error state.
Revision ChangesPath
1.36
mturk 2004/12/07 06:29:14
Modified:jk/native/common jk_lb_worker.c
Log:
Use static for local functions
Revision ChangesPath
1.35 +5 -5 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
==
mturk 2004/12/07 05:23:14
Modified:jk/native/common jk_lb_worker.c jk_uri_worker_map.c
jk_util.c
Log:
Fix compile time warnings
Revision ChangesPath
1.34 +2 -3 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_
mturk 2004/12/07 04:24:53
Modified:jk/native/common jk_lb_worker.c jk_lb_worker.h jk_util.c
jk_util.h
Log:
Make jk replication aware using domain model.
Great patch from Rainer Jung.
Revision ChangesPath
1.33 +219 -48 jakarta-tomcat-conn
mturk 2004/11/19 07:30:44
Modified:jk/native/common jk_lb_worker.c
Log:
Added configurable WAIT_BEFORE_RECOVER param.
The new directive is named "recover_time" and can be customized
inside workers.properies.
Revision ChangesPath
1.31 +15 -4 jakarta-tomcat-c
mturk 2004/11/11 10:43:07
Modified:jk/native/common jk_lb_worker.c
Log:
Add new trace macros and clean up some log messages.
Revision ChangesPath
1.29 +12 -5 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
mturk 2004/11/11 01:45:23
Modified:jk/native/common jk_lb_worker.c jk_msg_buff.c jk_msg_buff.h
jk_uri_worker_map.c jk_worker.c
Log:
Clean up log messages.
Revision ChangesPath
1.28 +4 -4 jakarta-tomcat-connectors/jk/native/common/jk_lb_
mturk 2004/11/08 05:42:23
Modified:jk/native/common jk_lb_worker.c
Log:
Backport load balancer algorithm from proxy_balancer.
Use integer balancer factors instead double.
This algorithim is much fatster and it actually behaves like
it should.
Revision ChangesPath
[EMAIL PROTECTED] wrote:
billbarker2004/07/27 22:59:44
Modified:jk/native/common jk_lb_worker.c
Log:
Check all JSESSIONID cookies for a valid jvmRoute.
If you have multiple Tomcats with overlapping domains, then you can get multiple cookies without a defined order. This will rou
billbarker2004/07/27 22:59:44
Modified:jk/native/common jk_lb_worker.c
Log:
Check all JSESSIONID cookies for a valid jvmRoute.
If you have multiple Tomcats with overlapping domains, then you can get multiple
cookies without a defined order. This will route correctly as long as
billbarker2004/07/20 19:18:46
Modified:jk/native/common jk_lb_worker.c
Log:
Fix nasty typo that was killing the cookie.
Reported By: Sandy McArthur <[EMAIL PROTECTED]>
Revision ChangesPath
1.22 +2 -2 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
hgomez 2004/07/20 07:56:40
Modified:jk/native/common jk_lb_worker.c
Log:
More debug infos around LB.
Revision ChangesPath
1.21 +8 -1 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
=
hgomez 2004/07/15 00:04:15
Modified:jk/native/common jk_lb_worker.c
Log:
Fix typo
Revision ChangesPath
1.20 +2 -2 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
billbarker2004/07/14 22:17:17
Modified:jk/native/common jk_lb_worker.c
Log:
Fix backwards test, and off-by-one on value.
Revision ChangesPath
1.19 +3 -2 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
==
billbarker2004/07/14 20:37:01
Modified:jk/native/common jk_lb_worker.c
Log:
Fix the parsing of the Cookie so that PSJSESSIONID doesn't match.
Also make the parsing a bit closer to RFC 2109
Reported By: Sandy McArthur <[EMAIL PROTECTED]>
Revision ChangesPath
1.18
hgomez 2004/02/06 00:37:59
Modified:jk/native/common jk_lb_worker.c
Log:
Comment end of LB service
Revision ChangesPath
1.15 +3 -3 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
===
glenn 2003/07/15 05:15:42
Modified:jk/native/common jk_lb_worker.c
Log:
Remove unnecessary error mesage when all lb workers fail
Revision ChangesPath
1.13 +2 -1 jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c
Index: jk_lb_worker.c
costin 02/05/15 14:00:29
Modified:jk/native build.xml
jk/native/common jk_lb_worker.c jk_util.c jk_util.h
Log:
Patch from Bernd Koecke, adding better support for hw balancers
Submitted by: Bernd Keocke [EMAIL PROTECTED]
Revision ChangesPath
1.33
costin 02/05/03 16:32:43
Modified:jk/native/common jk_lb_worker.c
Log:
Patch from Bernd Koecke, fixing the 'default worker'.
Note that a '0' value means the worker will get _all_ requests.
Revision ChangesPath
1.9 +13 -8 jakarta-tomcat-connectors/jk/nati
costin 02/01/11 21:27:39
Modified:jk/native build.xml
jk/native/common jk_lb_worker.c
Log:
Few changes to make it build on hpux.
A small fix in lb worker - try once more to connect to the host
where the session is located before giving up. This should solve
hgomez 01/06/18 01:32:22
Modified:jk/native/common jk_lb_worker.c
Log:
This patch changes the behavior by pre-initializing lb_value for each
worker. The selection algorithm searches for the worker with the lowest
lb_value that is not in a failed state. It then increments the
99 matches
Mail list logo
