HPOS 3.5 Solution for Unknown Status Occurring on Overloaded Servers

ManualsBrandsHP ManualsSoftwareHP Output Management Services Software

HP Output Server 3.5 1

Technical Note

Solution for Unknown Status Occurring

on Overloaded Servers

This technical note addresses the issue of some servers shutting down when overloaded. It also provides a

workaround for resolving the problem.

Servers affected

• Job Queue Manager (JQM) server

• Destination Manager (DSM) server

• Configuration Manager (CM) server

Problem

When a user attempts to execute an HP Output Server command on a batch of jobs, the JQM, CM and DSM servers

go into an ‘unknown’ status, or they generate a core dump. Executing an HP Output Server command on a batch of

jobs forces the servers to attempt to allocate more memory than allowed by the specified data segment stack size.

Solution

To avoid receiving an ‘unknown’ status, perform one of the following options:

• Increase the data segment stack size.

• Use indexed attributes for HP Output Server commands.

Increase the data segment stack size

Increase the data segment stack size by using the ulimit command. Each operating system uses a separate data

segment stack size. Find the current data segment stack size by using the following command:

ulimit -a

Example

Output for ulimit -a command on the Solaris platform:

$ulimit -a

Summary of content (6 pages)

PAGE 1
Technical Note Solution for Unknown Status Occurring on Overloaded Servers This technical note addresses the issue of some servers shutting down when overloaded. It also provides a workaround for resolving the problem.
PAGE 2
Solution for Unknown Status Occurring on Overloaded Servers Solution time (seconds) - unlimited file (blocks) - unlimited data (kbytes) - 96000 stack (kbytes) - 8192 coredump (blocks) - unlimited nofiles (descriptors) - 256 vmemory (kbytes) - unlimited NOTE: Check the help information of ulimit command on each operating system to change the data segment stack size.
PAGE 3
Solution for Unknown Status Occurring on Overloaded Servers Solution Server Name Privilege Manager (PM) Event manager (EM) CM Delivery Manager (DLM) HP Output Server 3.
PAGE 4
Solution for Unknown Status Occurring on Overloaded Servers HP-UX specific case Server Name JQM DSM Table Name Indexed Attribute 016 (dlm job) __name,current-dlm-job-state,dlm-job-submissi on-time 001 (job) __name,job-name,job-owner,destination,current -job-state,queue-assigned,job-completionstatus,physical-printer-assigned,job-rete ntion-time 002 (printer) __name,printer-realization,queue-supported 003 (server) __name 051 (queue) __name,scheduler-assigned 107 (capabilities) __name 002 (pr
PAGE 5
Solution for Unknown Status Occurring on Overloaded Servers HP-UX specific case Restarting JQM server causes the server to use a single ARENA for all the memory allocation requests preventing the server memory from increasing for any of the subsequent requests. NOTE: The memory growth for a specific server can also be restricted by setting the -server-environment-variables attribute for that server.
PAGE 6
Solution for Unknown Status Occurring on Overloaded Servers Conclusion This command forces the HP-UX heap manager to use a single ARENA. So when the CM server is run, the first pdconfig -d command invocation raises the memory size of the CM server from 14MB to 58MB while the subsequent pdconfig -d invocations do not raise the memory size above 58MB. NOTE: The use of the _M_ARENA_OPTS environment variable is described in the HP-UX malloc man page.