To Collect Required Debug Data For an Unresponsive or Hung Directory Server Process

This procedure describes what data to collect when Directory Server is still running, but is no longer responding to client application requests.

Collect the data describe in this procedure while the server is hanging.

1. Note the time during which the hang is seen to occur.

2. Collect information about the port used during the hang.

   UNIX and Linux

   netstat -an | grep ds-port

   Windows

   netstat -an

3. Collect statistics about the system running Directory Server.

   Solaris OS

   ps -aux | grep server-root

   vmstat 5 5

   iostat -x

   top

   uptime

   HP-UX

   ps -aux | grep server-root

   vmstat 5 5

   iostat -x

   top

   sar

   Red Hat

   ps -aux | grep server-root

   vmstat 5 5

   top

   uptime

   sar

   Windows

   Get the process ID using the tlist.exe command, then get process details using the same command.

   win-dbg-root\tlist.exe pid

4. Collect swap information.

   Solaris OS

   swap -l

   HP-UX

   swapinfo

   Red Hat

   free

   Windows

   Already provided in C:\report.txt.

5. On Solaris systems, collect output from pstack and pmap five times, once every ten seconds.

   pstack pid

   pmap -x pid

6. Get output showing system calls during the hang, by letting each of the commands listed here run for about a minute, then stopping them by typing Control-C.

   Solaris OS

   truss -ealf -rall -wall -vall -o /tmp/truss.out -p pid

   HP-UX

   tusc -v -fealT -rall -wall -o /tmp/truss.out -p pid

   Red Hat

   strace -fv -o /tmp/strace.out -p pid

   Windows

   Use Debug View.

   DebugView is available at http://www.sysinternals.com/Utilities/DebugView.html.

7. Collect core files or crash dumps, and related command output.

   When the server is hanging, attempt to get several core files that show the state of the server threads over time. You can do this by generating a core file, changing the name of the core file, waiting 30 seconds to a minute, and generating another core file. Repeat the process at least once to get a minimum of three sets of core files and related data.

   Solaris OS

   cd server-root/bin/slapd/server

   gcore -o /tmp/directory-core pid

   pstack /tmp/directory-core

   For each core file on Solaris OS, collect output from the following commands.

   cd server-root/bin/slapd/server

   file core-file

   pstack core-file

   pmap core-file

   pflags core-file

   For at least one of the core files on Solaris OS, collect output from the pkg_app script.

   ./pkg_app.ksh pid /tmp/directory-core

   Here, pid is the server process ID number. See To Run the pkg_app Script for instructions on using pkg_app.

   HP-UX

   cd server-root/bin/slapd/server

   If you have the patches PHKL_31876 and PHCO_32173 patches installed, generate the core file using the gcore command.

   gcore -p pid

   Otherwise, use the gdb command to generate the core file.

   gdb

   (gdb) attach pid

   Attaching to process pid

   No executable file name was specified

   (gdb) dumpcore

   Dumping core to the core file core.pid

   (gdb) quit

   The program is running. Quit anyway (and detach it)? (y or n) y

   Detaching from program: , process pid

   Red Hat

   cd server-root/bin/slapd/server

   gdb

   (gdb) attach pid

   Attaching to process pid

   No executable file was specified

   (gdb) gcore

   Saved corefile core.pid

   (gdb) backtrace

   (gdb) quit

   Windows

   win-dbg-root\tlist.exe

   win-dbg-root\adplus.vbs -hang -p pid -o C:\dump-dir

   Collect everything in the folder under C:\dump-dir.

8. Collect output from idsktune.

   The idsktune command provides information on system parameters, patch level, tuning recommendations, and so forth. The command is described in the product documentation.

   server-root/bin/slapd/server/idsktune