340 Chapter 5 Troubleshooting
Condition Recommended Action
Error in paired volume
operation
Hitachi TrueCopy only: If an error occurs in duplicated writing in paired volumes (i.e., pair
suspension), the server software using the volumes may detect the error by means of the
fence level of the paired volume. In such a case, check the error notification command or
syslog file to identify a failed paired volume.
The system administrator can confirm that duplicated writing in a paired volume is
suspended due to a failure and the system runs in regressed state using the error
notification command of the Hitachi TrueCopy. HORCM monitors failures in paired volumes
at regular intervals. When it detects a failure, it outputs it to the host’s syslog file. Thus, the
system administrator can detect the failure by checking the syslog file. Concerning the
operation of the RAID storage system, the failure can also be found on the Remote Console
PC (or SVP) provided.
Issue the Hitachi TrueCopy commands manually to the identified failed paired volume to try
to recover it. If the secondary volume is proved to be the failed volume, issue the pair
resynchronization command to recover it. If the primary volume fails, delete the paired
volume (pair splitting simplex) and use the secondary volume as the substitute volume.
About “/var(usr)/tmp” directory CCI uses “/var/tmp” or “/usr/tmp” as the directory for UNIX domain socket for IPC (Inter
Process Communication), and makes the directory and files as “/var/tmp/.lcm*” in CCI
version 01-16-06 or before.
Caution: This “/var/tmp/.lcm*” should not be removed while HORCM is running.
In case of Red Hat Linux, Cron executes the following “/etc/cron.daily/tmpwatch” file as
default:
------------------------------------------------------------
/usr/sbin/tmpwatch 240 /tmp
/usr/sbin/tmpwatch 720 /var/tmp
for d in /var/{cache/man,catman}/{cat?,X11R6/cat?,local/cat?}; do
if [ -d "$d" ]; then
/usr/sbin/tmpwatch -f 720 $d
fi
done
------------------------------------------------------------
The command of second line will remove “/var/tmp/.lcm*” directory after 720 Hr from
HORCM start-up, even though CCI command is used.
Action: So administrator needs to add the following command in order to avoid this problem:
------------------------------------------------------------
/bin/touch -c /var/tmp/.lcm* 2>/dev/null
/usr/sbin/tmpwatch 240 /tmp
/usr/sbin/tmpwatch 720 /var/tmp
for d in /var/{cache/man,catman}/{cat?,X11R6/cat?,local/cat?}; do
if [ -d "$d" ]; then
/usr/sbin/tmpwatch -f 720 $d
fi
done
------------------------------------------------------------
