Storage Resource Manager Enterprise Edition Installation Guide

Storage Resource Manager Enterprise Edition C–25
SRM Troubleshooting Guide
“Partition Could not be Scanned,” “Tree Scan Failed”
The Partition File scan on UNIX could not complete for some reason. Often this is due
to problems in the UNIX file system, and indicates a fsck should be run on the file
system to repair damage.
Consult the trace file created by the scan under /usr/srmagent/trace for this scan for
more details on why the scan failed.
Why aren’t Some of the Disks That are installed in My UNIX
System Visible in SRM?
If the disks are not assigned and visible directly to UNIX, the agent does not report
them to SRM. For example, if the disks are actually controlled by a built-in RAID
controller (like a DPT card), the disks actually seen by the system are those the
controller makes visible. In general, if the agent cannot find the size of a disk, it does
not report it. (This includes disks not “varied on” to AIX.)
“SRM was Unable to Establish a Connection to the UNIX Agent
Listening on Socket <ip_address>:<port>. The Connect
Request Failed With the Following Error 1006x”
Normally, you see the above message if the UNIX system is down. However, if SRM
cannot connect even when the system is up, it is normally due to a firewall
configuration issue. This is especially true if you can connect to the UNIX system
from the SRM system through “telnet” or “ping” it.
The “<port>” value in the message is the listening port on the UNIX system for the
SRM agent. If there is a firewall between the SRM server and the UNIX system, make
sure it allows TCP connection requests through on the port. (Or, re-install the SRM
Agent on the UNIX system on an allowed port.)
Also be sure that the UNIX Agent was configured to listen on either port 11125,
11126, or 11127. Any other port will fail.
“Too Many Concurrent Scans; Try Again Later.”
This is not an abnormal condition. There is a limit to the number of concurrent
partition scans which will run on a UNIX server, and it defaults to 3. Beyond this, a
“Busy” status is returned to SRM and it tries again later.