Level 2 Recovery; Level 0 Deadstart; Error Processing - Control Data Corporation CYBER 170 Series Operator's Manual

I
It is recommended that you stop current system activity
before beginning the system deadstart procedure (that is,
before activating the deadstart switch). To accomplish
this, enter the following DSD commands.
CHECK POINT SYSTEM.
Provides for termination of job processing and
writing the contents of central memory tables to
mass storage. For a complete description of this
process, refer to the CHECK POINT SYSTEM
com mand in section 3.
UNLOCK.
Necessary only if console is currently locked.
STEP.
Prevents the system from processing PP requests.
This stops all central memory I/O operations.
LEVEL 2 RECOVERY
Level
2
recovery deadstart is normally used in system test
situations and is not recommended for the normal
production environment. If level 2 recovery is selected, all
jobs and active files are recovered from checkpoint
information on mass storage as in level 1 recovery.
However, no attempt is made to recover the system.
Instead, the system is loaded from the deadstart file as in
level 0 deadstart. In all other respects, level 2 recovery is
identical to that described for level 1 in the preceding
description, and
all
rules apply.
LEVEL 0 DEADSTART
Level 0 or initial deadstart is used in cases where a
recovery deadstart is not possible. This is' a complete or
initial load from the deadstart file. Only preserved files,
which includes permanent files, queue files, and system
dayfiles, are recovered (preserved files are recovered on all
levels of system deadstart). Because memory confidence
testing destroys the contents of central memory and PPs,
all memory dumps must be completed before deadstart
begins.
If the machine is the first machine being
deadstarted in a muItimainframe
environment, the PRESET CMRDECK
entry (refer to Installation Handbook for
description) must be entered.
ERROR PROCESSING
If no display appears after activating the deadstart switch,
perform the following steps as needed. After each step,
activate the deadstart switch again to see if the problem
has been eliminated.
2-36
For deadstart from tape:
1.
If the unit select switch on the deadstart tape unit
is not on (tape does not move), check the channel,
controller, and .unit selections on the deadstart
panel to ensure they are set correctly.
2. If the unit select switch is on, the correct unit
was selected; however, check word 11 of the
deadstart panel to ensure it is set correctly.
3. Ensure that a 7-track tape is not mounted on a
9-track drive or vice versa. Also, ensure that a
deadstart tape with density of 6250 cpi is not
mounted on a tape unit which does not support
tha t density.
4. Ensure that the deadstart tape is an I-mode
unlabeled tape.
5. Ensure that the card reader and tape unit (667 or
669 only) are not on the same channel and the
card reader is not on a channel with a PP. Also,
ensure that two or more units do not have the
same physical unit number.
6. If still no display appears after activating the
deadstart switch, inform a site analyst. There
might be a parity error on one of the first records
of the deadstart tape or the magnetic tape
controller might have detected a channel' parity
error on a CYBER 170 Series machine.
For deadstart from disk:
1. Ensure that the disk is' spinning, the READY light
is on, and the SELECT light is on.
2.' Ensure that the disk has the CTI module loaded.
3. Ensure that the deadstart panel is set correctly.
4. Select an alternate channel.
5. If still no display appears after activating the
deadstart switch, inform a site analyst. There
might be a parity error on one of the first records
of the deadstart file or the disk controller might
have detected a channel parity error on a CYBER
170 Series machine.
For a proper understanding of the problems which may
occur during deadstart, there are several basic concepts
with which you should be familiar. For example, because
most errors that occur involve mass storage devices, you
should be familiar with their use in the system. Each mass
storage device has a' label that contains descriptive
·information about its contents. For certain levels of
recovery deadstart, this information must be consistent
with corresponding information contained in central
memory or provided through deadstart procedures.
Conflicts can result in deadstart error messages. An
attempt is made to recover
all
mass storage devices
defined in the EST during all levels of system deadstart.
The specific recovery function performed depends upon the
level of deadstart selected. Table 2-5 describes the
recovery function performed for each deadstart level as
well as the type of errors that can be encountered. The
system response to errors and the recommended action for
you to take are also listed.
Refer to appendix B for information concerning all
deadstart messages.
60435600 L