Fujitsu Siemens Computers CentricStor V3.1D, Руководство пользователя

Страница: 445 / 640

U41117-J-Z125-7-76
445
Explanation of console messages Message lines
9.2 Message lines
9.2.1 SXCF... (CMF: Cache Mirroring Feature)
SXCF012 CMF(#12): ALERT: failed to become PRIMARY, time exhausted: <sec> seconds
Meaning
During the switchover to multi-user mode (RunLevel 2) in the VLP’s boot phase the rc2
script
/etc/rc2.d/S70mirrorstart
determined that the data of the cache mirror when
obsolete when it checked it. The switchover to multi-user mode is stopped until the data
resources have been validated manually.
The cache mirror data ( /etc/mirrord/ms.db ) is replicated on the mirror backup server
(SVLP) by the mirror server (VLP) at regular intervals. It is regarded as obsolete after 500
seconds.
The time period is specified in the message. Depending on this period, different measures
must be taken.
Response
When a value greater than 500 seconds is specified as a period, a check must be made to
see whether CMF failures have been reported by the ICPs or IDPs since the time Time stamp
of the message minus period
. For this purpose messages which contain the
mirror chan-
ging string and have a corresponding time stamp must be searched for in the Klogger files
/var/adm/klog.msg* of these ISPs.
Case 1:
If no such message exists, the data resources are declared valid by the measure below.
Access to the VLP via LAN is possible despite incomplete multi-user mode as the LAN
services (inetd) have already been started.
The existing data resources can be validated using the following statement:
echo "reset -f" > /etc/mirrord/mirror_server_ctl
As an rsh call:
rsh <VLP-Name> "echo \"reset -f\" > /etc/mirrord/mirror_server_ctl"
The switchover to multi-user mode is then continued.
Case 2:
If messages of the type described above exist for the specified period,
/etc/mirrord/ms.db must be reconstructed manually. Notify 3rd-Level Support!