PDS in a loop

• Article Type: General
• Product: MetaLib
• Product Version: 3

Description:
Last night, our CRON job which runs UTIL-X-9 failed with the following error:
sort: can't map anonymous memory: Resource temporarily unavailable

We killed the job and restarted Metalib. Now, when we link to MetaLib , the system connects, but then keeps redirecting back and forth between PDS/SSO and Metalib.

The PDS server log is full of stuff like this:
2005-05-09 10:44:49 [PDSSso] RESP url = http://metalib.server.edu:80/V/VEJNQD5RILKL535P4EEIMQ87RQVLEKK5Y75FLNSSH669AXSF9K-00677?func=find-db-1&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GUEST&pds_handle=GU
2005-05-09 10:44:49 [PDSSso]

Resolution:
When the UTIL-X-9 cronjob fails and/or when vir01 is run without closing down MetaLib, the z54 session index is not properly reconstructed. Thus, when a user tried to access MetaLib, the system could not
assign a session id and got into a loop.
To resolve the problem, shut down MetaLib, run the clear_vir01 to completion (so session index is rebuilt) and then restart MetaLib in an orderly fashion.

• Article last edited: 10/8/2013