DEFINITY Enterprise Communications Server Release 5
Maintenance and Test for R5vs/si
555-230-123
Issue 1
April 1997
Maintenance Commands and Trouble-Clearing Aids
Page 8-159download update-file
8
Error on Application of the Patch
A patch may not have been applied for the following reasons:
1. The memory card was write-protected. Remove this protection and issue a
reset system x command
2. The patch identifiers were inconsistent. Run list configuration software
and compare the old_patch identifier with the values in the update file.
3. The LMM encountered a problem with the patch file. This is unlikely
because the same checks, and more, were performed when the file was
downloaded, prior to marking the file valid. This implies that the memory
which stored the update file was corrupted. Apply the back out file
immediately to back out the changes. Run the flash checksum test to
make sure the system is back to its prepatch state. Check the validity of
the file again with the development community and then try redownloading
and applying the patch immediately.
4. The LMM reports a hard error. The symptoms of this is an entry in the
hardware error log for the processor/memory board, if you’re lucky, or
extremely odd switch behavior followed by SPE down mode if you’re not.
The problem is that the LMM couldn’t complete the programming of
memory with the result that memory is in a corrupted state. The only
recovery is to visit the site armed with new software and processor/
memory circuit packs.
In a High or Critical Reliability System, the failure causes a switch to the
standby processor. The hardware on the standby must be repaired and
the patch redownloaded. (There was nothing wrong with the patch)
Good Application - Bad Patch
This error is caused, not by a failure in the download or application, but by a fault
in the patch file itself. To recover from this type of problem, the back out file
which backs out the patch should be downloaded and applied. Clearly, this
requires that the system be sane enough to receive the file correctly and be able
to apply it.
In a High or Critical Reliability System, the user has about eight minutes to
recognize that a problem exists and force an interchange to the standby
processor. If this can be done, the file on the newly active processor can be
invalidated using a file containing a destroy tuple or the
wp byte
command. The
standby can be restored to a normal state using the back out file.
Inconsistent Software Versions on a Duplicated
Switch
Inconsistent software, as indicated by a failure in the data consistency test, can
be caused by problems copying the update file to the standby or validation test
failures on the standby. Unlike the tape or MIPS systems which revert to the same
version of software as a result of a refresh, a flash system remains inconsistent