Open GoogleCodeExporter opened 9 years ago
martin.grotzke:
hi,I'm urgenly waiting for your help.
and I tried issue 151, still can't work.
Original comment by zheng.gu...@gmail.com
on 11 Nov 2014 at 8:01
Where do you read the value? Are you sure that this happens after the ajax
request was completed?
Original comment by martin.grotzke
on 11 Nov 2014 at 8:12
[deleted comment]
[deleted comment]
I think I'm using msm in a very common context and should not meet these
problems.
maybe any jar version is something wrong ?
Original comment by zheng.gu...@gmail.com
on 11 Nov 2014 at 9:59
Can you activate debug logging according to the docs and share the log output?
Original comment by martin.grotzke
on 11 Nov 2014 at 10:01
[deleted comment]
Do you have the related tomcat logs (with debug mode enabled)?
Original comment by martin.grotzke
on 12 Nov 2014 at 8:15
[deleted comment]
[deleted comment]
[deleted comment]
[deleted comment]
[deleted comment]
[deleted comment]
[deleted comment]
[deleted comment]
log file as attachment .
/wmt365/user!randomImage.action create random and set to session attribute.
/wmt365/login!login.action check request random with session attribute random
I tested twice .
the first is ok,the second is bad.
And I saw that ,the first time ,randomImage.action set attribute to the
session,and when logined,
"2014-11-21 00:36:44,257 [http-bio-80-exec-10] DEBUG
de.javakaffee.web.msm.SessionIdFormat- Creating new session id with orig id
'ping' and memcached id 'n1'.
2014-11-21 00:36:44,260 [http-bio-80-exec-10] DEBUG
de.javakaffee.web.msm.NodeAvailabilityCache- CacheLoader returned node
availability 'true' for node 'n1'.
2014-11-21 00:36:44,261 [http-bio-80-exec-10] DEBUG
de.javakaffee.web.msm.MemcachedBackupSessionManager$1- Loading session from
memcached: F05F8770BB6E6CE2BB968C713088F840-n1"
Yes, session is loaded from memcached.
but the second time:
2014-11-21 00:36:50,449 [http-bio-80-exec-1] DEBUG
de.javakaffee.web.msm.SessionIdFormat- Creating new session id with orig id
'ping' and memcached id 'n1'.
2014-11-21 00:36:50,452 [http-bio-80-exec-1] DEBUG
de.javakaffee.web.msm.NodeAvailabilityCache- CacheLoader returned node
availability 'true' for node 'n1'.
2014-11-21 00:36:50,460 [http-bio-80-exec-1] DEBUG
de.javakaffee.web.msm.MemcachedBackupSessionManager$1- createSession invoked:
F05F8770BB6E6CE2BB968C713088F840-n1
2014-11-21 00:36:50,460 [http-bio-80-exec-1] DEBUG
de.javakaffee.web.msm.MemcachedBackupSessionManager$1- Created new session with
id F05F8770BB6E6CE2BB968C713088F840-n1
2014-11-21 00:36:50,570 [http-bio-80-exec-1] DEBUG
de.javakaffee.web.msm.BackupSessionService- Starting for session id
F05F8770BB6E6CE2BB968C713088F840-n1
"Created new session" ,why created again ? not loaded from memached?
Can you help me out ? Thanks a lot.
Original comment by zheng.gu...@gmail.com
on 20 Nov 2014 at 4:47
Attachments:
martin.grotzke:
Hello. I found the problem about the LRUCache.
====
In MemcachedSessionService.java :
private final LRUCache<String, Boolean> _invalidSessionsCache = new LRUCache<String, Boolean>( 2000, 5000 );
====
you defined the cache time to 5000,that's 5 seconds long.
When I logouted ,no sticky session called removed.
MemcachedBackupSessionManager- remove invoked, removeFromMemcached: true
thus ,the sessionid is input to the _invalidSessionsCache.
in short time(less than 5 seconds) after random created and input to session
attribute,I relogined, but the session id was indicated to invalid one.
then recreated the session ,and the random session attribute was also dispeared.
Thus, I have two ways :
1: reduce the LRUCache timeout from 5000 to 500 or more less ?
2: like another issue, remove the cookie ,every session with different cookie
Original comment by zheng.gu...@gmail.com
on 21 Nov 2014 at 6:54
Original issue reported on code.google.com by
zheng.gu...@gmail.com
on 11 Nov 2014 at 6:50