Thursday, December 29, 2016

What are the Content Server Tracing Options


Below are some of the tracing options supported by the Content Server:




ca_store_trace           Traces the Centera code

workflow_agent_trace     Traces actions of the workflow agent

file_store_trace         Traces reading/writing to a normal filestore

dql_update_trace         Trace DQL update statement processing

rpctrace                 Traces all RPC calls made by all DFC clients

dqltrace                 Traces DQL statements issued to the Content Server

sql_perf_trace           For Oracle, gives timing information for SQL statements

group_cache_trace        Traces group cache building and dynamic groups

trace_http_post          Traces communication between the JMS and CS

crypto_trace             Traces dmcryptokeymgr.cxx

acs_connector_trace       Traces ACS and CS communication

docbroker_trace           Traces CS to docbroker communication

log_authentication_errors Identical to trace_authentication

trace_authentication      Traces authentication functionality

ticket_trace              Traces login ticket allocation, and authentication

net_ip_addr               Prints out some hostname IP Addresses

stack_walk_off            Will force CS NOT to walk to stack on an exception

development               Debugging asserts - used for development

assert_fatal              Will the CS die during an assert - used for development

ecpool_trace              Traces dmExecutionContextPool

nettrace_all              Traces netwise layer

nettrace                  Traces netwise layer


audit_purge_policy_cache_trace     Tracing of dmSession::ValidateAuditPurgePolicyCaches,     dmUpdateAuditPurgePolicyChangeTime, and dmCompareAuditPurgePolicyChangeTime



How to enable tracing:



There are two options to enable and disable Content Server tracing.



(1)Startup Option (Requires Content Server restart to take effect)


Unix: Example: dm_shutdown_cs67 -orpctrace -osqltrace
Windows: Use Documentum Server Manager to edit the service and add tracing (example: -orpctrace -osqltrace).

The tracing will be active until the corresponding startup option has been removed and the Content Server has been restarted.



(2)iAPI Options (No Content Server restart required to take effect)



#Example commands to enable:



API> apply,c,NULL,SET_OPTIONS,OPTION,S,sqltrace,VALUE,B,T

API> apply,c,NULL,SET_OPTIONS,OPTION,S,docbroker_trace,VALUE,B,T

API> apply,c,NULL,SET_OPTIONS,OPTION,S,rpctrace,VALUE,B,T

API> apply,c,NULL,SET_OPTIONS,OPTION,S,ticket_trace,VALUE,B,T

API> apply,c,NULL,SET_OPTIONS,OPTION,S,trace_authentication,VALUE,B,T



It would be advisable to do API> reinit,c after each command to ensure the tracing has been activated. It may also be recommended to verify the logs immediately after entering the command to ensure they contain the relevant level of tracing.




#Examples commands to disable:



API> apply,c,NULL,SET_OPTIONS,OPTION,S,sqltrace,VALUE,B,F

API> apply,c,NULL,SET_OPTIONS,OPTION,S,docbroker_trace,VALUE,B,F

API> apply,c,NULL,SET_OPTIONS,OPTION,S,rpctrace,VALUE,B,F

API> apply,c,NULL,SET_OPTIONS,OPTION,S,ticket_trace,VALUE,B,F

API> apply,c,NULL,SET_OPTIONS,OPTION,S,trace_authentication,VALUE,B,F



Again, after each CMD, do API> reinit,c to deactivate the tracing.



NOTES:


Some of the above tracing options can have a certain performance impact on the Content Server and should be applied with caution in production environments. Normally, they will be applied only at request of EMC Support/Engineering and for the only purpose of troubleshooting a specific issue. Ideally, the tracing should be enabled just prior to reproducing the issue under investigation and disabled immediately after the test has been completed. Also, where possible, the corresponding tracing should be collected in production environments outside working hours.

Wednesday, December 14, 2016

DM_METHOD_E_INVALID_MTHD_VERB

Symptoms
[DM_METHOD_E_INVALID_MTHD_VERB]error: "The dm_method named (dmclean) of type dmbasic has invalid method_verb (./dmclean)."

Cause
Apparently, there is a bug with the UI in DA. The method_type attribute is supposed to be set to '  '  for dmclean method.

Resolution
Update method_type of dmclean method to `  `. This will resolve the issue.

query : select * from dm_method where object_name='dmclean'