This section is a reference for the procedures used when diagnostics
are required.
Boot Modes
The SNC has two modes of operation: normal mode and diagnostics mode. Two
commands are available from the service terminal to reboot the SNC to the
desired mode.
The SNC remains in the existing mode until it has been directed to reboot in
the other mode.
When the SNC is booted in diagnostic mode, the command prompt displayed on
the service terminal is diagmode >. While in
diagmode, a limited command set is available for testing the SNC's interfaces.
In addition, the device drivers for the Ethernet, FC, and SCSI interfaces
are disabled so loop back tests can be performed on them.
Entering Diagnostic Mode
Use the diagBoot command when the special features
of the SNC 's diagnostic module are required. See diagBoot.
Perform the following steps:
Disconnect all SCSI and FC cables from the SNC. If you do not, the data
on the attached disk or tape devices will become corrupted.
Perform a proper shutdown of the tape library.
Remove the Ethernet cable, if installed. See Connecting
to the Service Port regarding further information for Step 3 and Step
4.
Connect the terminal to the service port.
Turn on the tape library. The tape library takes longer to boot than the
SNC.
Do one of the following:
If the command prompt displayed on the service terminal is
diagmode >, proceed with the appropriate diagnostic tests.
If the command prompt displayed on the service terminal is not diagmode
>, type the command diagBoot.
Wait for the SNC to reboot.
When the diagmode > prompt appears, proceed
with the appropriate diagnostic tests.
Restoring Normal Mode
Perform normalBoot if the SNC was placed
in diagnostic mode and needs to be restored to normal operation. See normalBoot.
From the service terminal, type the normalBoot
command.
Wait for the SNC to reboot.
At the prompt, continue with the MAP that sent you here.
Special Procedures
This section contains information on special procedures.
Health Check
The health check function provides a method to periodically determine the
operational state of an SNC and its attached interfaces and devices.
Manual Health Check
Health check is performed manually from the service terminal only if the SNC
is in normal mode. This procedure uses the hlthChkNow
command that performs a "device ready" level health check that includes a SCSI
device availability check. The SCSI target devices must be connected and turned
on so that health check can determine their operational status.
For other commands associated with health checks, see terms beginning with
the hlthChk prefix in
Service Port Command Reference.
Stop all I/O.
Remove the Ethernet cable if installed.
Connect the terminal to the service port.
From the service terminal, type the hlthChkNow
command.
Check the results displayed on the service terminal to determine the status
of the interfaces and attached devices.
If a failure is indicated, perform the appropriate MAP.
Event Log Dump
The events that have been logged in the SNC 's event log can be displayed
on the service terminal. This information may be
useful because you can see the normal events and any abnormal events that may
lead you to the source of an internal or external failure.
Each entry contains a sequence number, day, and time-stamp, the event code,
event source (interface name or system process), an index, and a description
of the event. If the real time clock has been set, the time-stamp will be real
time. Otherwise, the time stamp will be relative to the boot-time.
The event log holds more than 2000 events. When the log becomes full, the
oldest events are overwritten by the most recent events. You can choose to view
only those events logged since the SNC was last booted or you can view a longer
list of events that have accumulated from previous boots. You can choose to
view events by severity level. Selecting Warning will display
warning and notice events. Selecting Information will display
information, warning, and notice events.
Two kinds of events are recorded in the command flow log file: sequence events
and single command events. Sequence events are Read/Write commands. All other
events are single command events. Several commands can be used to produce command
flow log information:
cfLogTimePlayback
cfLogTimePlaybackAll
cfLogEventPlayback
cfLogEventPlaybackAll
Use these commands to get command flow log information limited by a particular
LUN, time period, or a specified number of events.
The output that results from these commands starts with the internal LUN number
of the device for which information is being provided. The device’s vendor,
product ID, and serial number are also listed in this line:
After this line, numbered events pertaining to the cited LUN are printed. Each
line begins with an event number. In the following example, the event number
is 506:
The event number is followed by the month, day of the month, start time, initiator
channel and initiator index. In event 506, the word "Write" after
the initiator index indicates that the sequence event "Write" is being
logged. The number "021540" after "Write" represents the
number of times the SNC carried out the "Write" command. No errors
were recorded in the course of the command being carried out. Compare the time
stamp at the end of this line to the start time to determine how long it took
the SNC to complete the sequence event.
If data was transferred to or from the host, a new line is displayed indicating
the direction of the transfer ("DIR=IN" is to the host and "DIR=OUT"
is from the host), the number of bytes requested (REQ) and the number of bytes
actually transferred (XFER):
DIR=IN REQ=14 XFER=14
If the event is a single command event, as in event 511 below, the initiator
index is followed by an 8 character abbreviation for the operation code, the
command descriptor block—which may be 6, 10, 12 or 16 bytes in length—the
command completion status, and end time:
The status origin is shown as DEV if the status is from the device and SNC
if the SNC returned the status. The sense data fields for Filemark, EOM, ILI,
key, additional sense code, and additional sense code qualifier are translated
and displayed on the next line. Refer to a SCSI command reference guide to understand
the sense data in these fields.
This event is also recorded in the SNC event log along with the name
of a file that was saved to the flash file system. The file contains information
about the state of the FC interface prior to the reset. Follow these
steps to retrieve the file.
Press the <Enter> key on the service terminal. If the prompt is not
displayed, go to Service Port MAP to determine
whether the RS-232 cable and service terminal are working properly.
From the service terminal, navigate to the DUMP
directory by entering the cd /ffs/dump command.
From the service terminal, enter thels command
to get a file listing for the directory. There may be more than one file present
if the error condition has occurred more than once.
NOTE: Dump filenames end in
.dmpand the files
are automatically deleted whenever you update the SNC's firmware.
Save files with the communications software you are using or see Steps
6 - 8 for HyperTerminal instructions. Go to Step 9 when the files are saved.
When using a HyperTerminal window, use sz fc1.dmp
The filename fc1.dmp is used as an example.
Specify a filename between quotes, with 8 characters or fewer plus the
file extension .dmp.
From the HyperTerminal window, select Transfer and Receive File.
Repeat Steps 6 and 7 for additional files when using a HyperTerminal window.
Go to Step 9.
Contact Dell for information about how to send the dump file to Dell for
analysis. See Getting Help.
Boot Mode Commands
diagBoot
This command is only used to transition an SNC from normal operations to the
special diagnostic mode and does the following:
Ensures that the /nvfs/mt directory exists
Copies the existing bootline to a file in the /nvfs/mt
directory on the SNC
Installs a new bootline directing the SNC to boot using a special diagnostic
startup script /ffs/mt/diagnstk.rc
Renames the persistent map file config/device.map
as config/device.bak (a new file will be generated
after rebooting)
Issues a reboot command to put the changes into effect
normalBoot
The normalBoot command is only used to transition
the SNC from the special diagnostic mode to resume normal operation and does
the following:
restores the bootline that was copied by diagBoot
erases the new persistent device map and renames the original map file
config/device.map - restoring it for use when the SNC reboots
reboots the SNC
Diagnostic Commands
This section is a reference for the commands that are available when the SNC
is booted in diagnostics mode.
The command set is limited while in diagmode. You will need to use the
showBox command and three commands with the suffix
Test when testing the Ethernet, FC, and SCSI interfaces.
An Ethernet loop back plug, FC loop back plug, and SCSI cable
are required for the corresponding loop back test.
NOTICE:
You must disconnect all SCSI and FC cables from the SNC. If you do not,
the data on the attached disk or tape devices will become corrupted.
elTest
elTest is the Ethernet loop back test. The SNC
must be in diagnostic mode to use this command and a loop back plug must be
installed on the Ethernet port.
The test issues a series of loop back tests. Test data is transferred and
verified. A good test ends with Ethernet OK.
If errors are detected, the test displays the number of bad test iterations
as shown below. If your test was not successful, return to the MAP that sent
you here.
fcSlotTest [x]
The SNC must be in diagnostic mode to use this command and a loop back plug
must be connected to the GBIC.
The fcSlotTest command performs a confidence
test on an FC slot. Substitute the FC slot number you want to test for x. This
command can be used to test the slot or the cable.
The following example code shows the display for a test for FC slot 1.
Unsuccessful Fibre Test 1
NOTE:
If the FC "link" LED does not come on, the test will not pass.
scsiChannelTest (x,y)
The SNC must be in diagnostic mode to use this command and a SCSI cable
must be connected between two SCSI channels.
The scsiChannelTest command performs a confidence
test on a pair of SCSI channels. Substitute the SCSI channel numbers you want
to test for x and y. Parenthesis are optional. This command
can be used to test the SCSI interface or the SCSI cable.
The example below shows the display for a test on SCSI channels 1 and 2.
The following output is an example of a SCSI port test that was not successful.
If your test was not successful, return to the MAP that sent you here.
Unsuccessful SCSI Test (Cable Not Plugged In)
showBox
The showBox command displays the
components present in the SNC using characters to form a picture of the unit,
as viewed from the back. This command also lists the cabling requirements for
each slot.
Figure 1. showBox
supportDump
The supportDump command outputs the result of a number
of service port commands in a standard way and in a specific order. The resulting
file is used for troubleshooting. The individual commands called by supportDump
are:
