HP XC System Software Release Notes for Version 3.2.1

13.2.1 Myrinet Monitoring Line Card Can Become Unresponsive

A Myrinet monitoring line card can become unresponsive some period of time after it has been

set up with an IP address with DHCP. This is a problem known to Myricom. For more information,

see the following:

http://www.myri.com/fom-serve/cache/321.html

If the line card becomes unresponsive, re-seat the line card by sliding it out of its chassis slot and

then slide it back in. You can do this while the system is up; doing so does not interfere with

Myrinet traffic.

13.2.2 The clear_counters Command Does Not Work On The 256 Port Switch

The /opt/gm/sbin/clear_counters command does not clear the counters on the Myrinet

256 port switch. The web interface to the Myrinet 256 port switch has changed from the earlier,

smaller switches.

To clear the switch counters, you must open an interactive Web connection to the switch and

clear the counters using the menu commands. The gm_prodmode_mon script, which uses the

clear_counters command, will not clear the counters periodically, as it does on the smaller

switches.

This problem will be resolved in a future software update from Myricom.

13.3 QsNet

Interconnect

The following release notes are specific to the QsNet

® interconnect.

13.3.1 Possible Conflict with Use of SIGUSR2

The Quadrics QsNet

software internally uses SIGUSR2 to manage the interconnect. This can

conflict with any user applications that use SIGUSR2, including for debugger use.

To work around this conflict, set the environment variable LIBELAN4_TRAPSIG for the application

to a different signal number other than the default value 12 that corresponds to SIGUSR2. Doing

this instructs the Quadrics software to use the new signal number, and SIGUSR2 can be once

again used by the application. Signal numbers are defined in the /usr/include/asm/signal.h

file.

13.3.2 The qsnet Database Might Contain Entries to Nonexistent Switch Modules

Depending on the system topology, the qsnet diagnostics database might contain entries to

nonexistent switches.

This issue is manifested as errors reported by the /usr/bin/qsctrl utility similar to the

following:

# qsctrl

qsctrl: failed to initialise module QR0N03: no such module (-7)

In the previous example, the switch_modules table in the qsnet database is populated with

QR0N03 even though the QR0N03 module is not physically present. This problem has been

reported to Quadrics, Ltd.

To work around this problem, delete the QR0N03 entry (and any other nonexistent switch entries)

from the switch_modules table, and restart the swmlogger service:

# mysql -u root -p qsnet

mysql> delete from switch_modules where name="QR0N03";

mysql> quit

# service swm restart

42 Interconnects