In that case this would be marked as 'notabug' [:)]
Well, no; the related Mantis issue reads:
13608: In a cluster-environment, the behavior of the list of webadmin sessions is undefined.
That aspect still needs to be checked, and that's why this thread is still marked CONFIRMED/OPEN. It is probably a rather small issue and not terribly urgent, but we should definitely not forget about it.
Now i am facing new issues when i have 11 failed login at bottam the sr no doesnt change it remain at 33 ,to chack i logout and login with incorrect passwd then i login with correct login ,correct login was get noticed by login session but failed login was not get notice ,fail login suppose to be 2nd number but that is also Successful
so some login session are not get registered !
some time FAILED LOGIN SESSION SHOW FAILED WHEN U LOGIN AFTER SOME LOGINS /LOGOUT sessions edit::::[:D]o u think this is one of the problem "while u enter user name and passwd some time after passing 36% or 90 % webadmin login process stall doesn't go through /webadmin page not display ,login process stall " thanks
13608: In a cluster-environment, the behavior of the list of webadmin sessions is undefined.
That aspect still needs to be checked, and that's why this thread is still marked CONFIRMED/OPEN. It is probably a rather small issue and not terribly urgent, but we should definitely not forget about it.
Here is what i found when looking into this:
Even though the function userlog_read() was not designed for cluster/HA -
it turns out it just works.
The log file format is basically the same, except that the host name carries a node suffix. As the host name is ignored by the log file parser, that's no problem. The log file parser is using the log file entries from all nodes. For WebAdmin purposes, only those written by the webadmin and acc_agent facilities are used. On the other nodes, the relevant entries are written by the system facility, user ha_sync, so they don't get in the way.
13608: In a cluster-environment, the behavior of the list of webadmin sessions is undefined.
That aspect still needs to be checked, and that's why this thread is still marked CONFIRMED/OPEN. It is probably a rather small issue and not terribly urgent, but we should definitely not forget about it.
Here is what i found when looking into this:
Even though the function userlog_read() was not designed for cluster/HA -
it turns out it just works.
The log file format is basically the same, except that the host name carries a node suffix. As the host name is ignored by the log file parser, that's no problem. The log file parser is using the log file entries from all nodes. For WebAdmin purposes, only those written by the webadmin and acc_agent facilities are used. On the other nodes, the relevant entries are written by the system facility, user ha_sync, so they don't get in the way.
