Uproc Aborts with unibatchenv authentication failed
Article ID: 249904
Updated On:
Products
Issue/Introduction
Randomly, a Uproc Aborts as it fails to be submitted as the unibatchenv process cannot load the environment.
The following errors appear in universe.log at the time:
|ERROR|X|IO |pid=p.t| u_io_thread_trt | New client XXX (unibatchenv/ on server.domain) authentication failed: ()
|ERROR|X|DQL|pid=p.t| k_handshakeAuthent | u_req_serv to [SERVER_local_SIO_X]/[XXX] in error [-2]
|ERROR|X|DQL|pid=p.t| k_connect_auth | Request authentication to [SERVER_local_SIO_X]/[XXX] in error [-1] (check parameter timeout for UVMS connexion )
|ERROR|X|DQL|pid=p.t| o_connect_auth | k_connect_auth_timeout returns error [205]
This occurs during a period where the CPU load of the server was quite high, relaunching the Uproc later on works fine.
Environment
Release : 6.x
Component : DOLLAR UNIVERSE
OS: AIX
Cause
Root cause unknown, one-time only issue most likely caused by high load on the uxioserv process which did not authorize the request from unibatchenv on time (10 seconds).
Resolution
Workaround:
Relaunch the affected Uproc/s
Long term Workaround:
Increase the variable IO timeout setting (U_IO_TIMEOUT) from 10 to 60 in node settings - technical setings
Then restart the node to take into account the modification.
Additional Information
Should the issue still occur after the change above, please open a case with Technical Support for further investigation.
Feedback
