
PROBLEM: (90020) (PATCH ID: TCR520-029)
********
When CAA took a long time to process certain EVM events, it would delay
actions taken in response to other events. This limitation has been removed
by increasing parallelism.


PROBLEM: (90159, ) (PATCH ID: TCR520-035)
********
The problem causes core dumps for every time resources are started and stopped.
The resource move to the unknown state and must be stopped with the -f option
and unregistered.


PROBLEM: (89445, 89446, 89474, 89475,
          89543, 89688, 89936, 90026) (PATCH ID: TCR520-022)
********
Enablers for Huron Packaged Database Solution


PROBLEM: (90123) (PATCH ID: TCR520-032)
********
The CAA datastore may become corrupted due to improper datastore locking. This
may occur when multiple CLI commands are run in the background from the same
member. The problem does not appear to arise when commands are run sequentially.


PROBLEM: (89449) (PATCH ID: TCR520-054)
********
When manipulating CAA resources with names that contain characters other than
alphanumeric and '_', caa_profile command may output messages similar to:
Failed to create event 'sys.unix.clu.caa.app.profile._name.xx-yy'
Error: Invalid value
Evm log has failed.
This update corrects this problem.


PROBLEM: (90311, UI) (PATCH ID: TCR520-047)
********
When a user uses "caa_profile -create" command to create a resource profile
for application resource, one extra line for "REBALANCE" is always being
inserted into the profile.  This will cause the CAA GUI to fail to validate
the profile and report that there is an error.


PROBLEM: (86965) (PATCH ID: TCR520-048)
********
All CAA CLI commands may crash when they are interrupted by SIGPIPE.  A core
file will be left in root directory when this occurs.  This problem has been
seen to occur when caa_stat output is piped to a command like "head".


PROBLEM: (90448) (PATCH ID: TCR520-051)
********
CAA does not properly reclaim space in the CAA registry when unregistering
resources with long resource or attribute names.


PROBLEM: (90607) (PATCH ID: TCR520-056)
********
Enabler for Huron Packaged Database Solution


PROBLEM: (90223) (PATCH ID: TCR520-046)
********
Repeated uses of caa_stat -f will cause the caad daemon to leak memory.
Eventually, it will terminate and dump core when its data ulimit is reached.


PROBLEM: (90392) (PATCH ID: TCR520-052)
********
CAA fails to close a TDF after processing a corresponding resource profile.
Over time this will lead to reaching the process limit for open file
descriptors and will prevent CAA from functioning properly.


PROBLEM: (89919) (PATCH ID: TCR520-049)
********
If the Event Manager daemon (evmd) terminates temporarily, the clu_mibs agent
currently attempts to reconnect to the daemon for up to 5 minutes, and then
terminates.  With this patch applied the agent continues to retry indefinitely,
and resumes normal service once the Event Manager daemon is restarted.
This patch also improves user control of the SNMP service by providing start
and stop control through /sbin/init.d/snmpd script.


PROBLEM: (90676, 81679) (PATCH ID: TCR520-065)
********
There are several problems that this patch fixes.
PROBLEM (): (PATCH ID:)

* Create 2 resources, A, B, where A requires resource B.
  A's script returns 1 on start, indicating a failure. Start B "caa_start B" 
  and then start A, "caa_start -f A"
  The result is that A does not start and B will be stopped also.
  B should not stop, the patch resolves this problem.

* Create 3 resources, A, B, C, where A requires resource B, and C requires
  resource B. A's script returns 1 on start, indicating a failure.
  Start A, and B will start also. The start script causes a failure condition,
  so resources A and B are in the online/offline state. Then start C. 
  This fails since it trys to start A and A fails, and C fails to start.
  The patch fixes this since CAA should not start A it is not online. 

* Create 3 resources, A, B, C, where A requires resource B and C requires
  resource B. A's script returns failure for both starts and stops. Start
  A and resource A moves to the offline/Unknown state, and try to start C.
  C will fail to start.

* Create 3 resources, A, B, C, where A requires resource B, and C requires
  resource B. Register A and start A. Resource A goes to the Unknown state.
  Then register C. This results in a core dump.

* Create 2 resources A, B, where A requires resource B. Start A on one node.
  Resource A and B are started. Then relocate "caa_relocate -s -c othernode".
  Relocate all of the resources from node X to node Y. This fails.

* Create 2 resources A and B, where A requires B. They are started on node X.
  Relocate with the -f flag to another node. The caa_start's return value is 
  incorrect, with value of "223" and the messages
  indicate that the placement had an error, which the results are successful.
  The patch  resolves this problem.

* Create 2 resources A and B, where A requires B. Start the resources, and
  relocate using the -f flag where, "caa_relocate -f B A". Notice how
  both resources are stopped and srestarted on the new node, but
  then they are stopped and restarted back to the original node.


PROBLEM: (90729) (PATCH ID: TCR520-060)
********
Enabler for Huron Packaged Database Solution


PROBLEM: (90834) (PATCH ID: TCR520-063)
********
Enabler for Huron Packaged Database Solution


PROBLEM: (90638) (PATCH ID: TCR520-072)
********
The new caads running on the rolled members in the cluster roll utilize
a new datastore. The information in the old datastore was never copied
to the new datastore. After all the members have rolled the default
resources seen in the old datastore will not be found.


PROBLEM: (91262) (PATCH ID: TCR520-102)
********
While all known problems that can cause the CAA deamon to fail during
initialization have been corrected, this fix ensures that if any such problems
are experienced, the CAA boot script will exit with an error instead of
hanging the boot process.


PROBLEM: (90977, 90978, 90979) (PATCH ID: TCR520-071)
********
PROBLEM: 90977
The CAA caad daemon engine will core dump when trying to restart
on one cluster node while the caad daemon engines on other cluster
nodes are up with heavy CAA activity.  
PROBLEM: 90978
There is an excessive delay encountered while trying to restart 
the CAA caad daemon engine on one cluster node while the caad daemon
engines on other cluster nodes are up with heavy CAA activity.
PROBLEM: 90979
There is an unexpected silent termination of the CAA caad daemon
engine under conditions of heavy CAA activity.


PROBLEM: (90855, 94755) (PATCH ID: TCR520-201)
********
Because of an ordering problem in one of the rc scripts
shipped with TruClusters, Oracle can fail during start-up.


PROBLEM: (91152, FR_G06068) (PATCH ID: TCR520-241)
********
The application resource does not go offline even when its last network
dependent resource is gone.


PROBLEM: (95782) (PATCH ID: TCR520-249)
********
PROBLEM:
CAAD might coredump due to a race condition when multiple events to which it
subscribes arrive at the same time when CAAD has not seen any events after
startup. This problem is seen mostly when the cluster loses Quorum due to a
member leaving the cluster.


PROBLEM: (BCSMP0BVJ) (PATCH ID: TCR520-219)
********
This patch fixes the problem when the target member crashes during service
start up.
The problem manifests as the CAA services are in an undefined state where they 
cannot be stopped, started, or relocated. A system reboot or manually restarting
the CAADs are needed to correct the problem.


PROBLEM: (95805) (PATCH ID: TCR520-266)
********
PROBLEM:
        Currently the startup and failover sequence for services in CAA is done
in the alphbetical order of the service names. This created a problem where in a
user service having an implict dependency on a system service could not start on
a node if it was started-up or failed over before the system service became
available because of its name being higher up in the order.
SOLUTON:
        This problem is fixed by committing to a pioritized failover sequence
where in a service designated as a system service is always started or failed
over before any of the user services.



