"Eric Thomas (CERN/L3)" <ERIC@LEPICS>
Thu, 17 Nov 88 22:32:02 GMT
|
Some tests I (painfully) managed to make on LISTSERV@AEARN, which was a very
convenient server to make tests on since it's directly connected to CEARN over
a lightly loaded line, does not have any list and is not on the backbone,
showed that the error 32 from CARD LOAD is something temporary. By erasing
INSTALL FILE and re-sending the update files, I was able to get the code to
install successfully without doing anything special. It sounds like the first
set of files arrived corrupted on the reader of [log in to unmask]
Now the problem is that I cannot resend the shipment before INSTALL FILE has
been modified to get rid of the "failure" information. There is a command,
INSTALL CLEANUP, which would do that, but unfortunately it doesn't work if the
shipment was not installed (so that you can't inadvertently discard a shipment
which is still awaiting segments). This means you'll have to either TELL
LISTSERV CMS ERASE INSTALL FILE, or XEDIT INSTALL FILE and delete all the
lines which refer to CODE_1.5O. When you have done that, just send me a note
and I'll resend the files (I plan to batch as many sites as possible together,
to save network bandwidth).
Now I'd like to know how the files got corrupted in the first place. All the
nodes past UMSLVMA seem to have been affected, along with AEARN and a few
nodes in Canada. This is particularly strange since AEARN got its copy
directly from CEARN, which distributed ALL the copies and can't be suspected
of having sent a garbled file (or then only one of the output copies was
affected, which seems strange). In the case of UMSLVMA et seq, it can be
assumed that some node along the path clobbered the file and that the
following LISTSERVs just distributed the belaboured copy onwards. Any
information you could provide about this would be welcome - it's not the first
time CARD decks get maimed, and it probably won't be the last time if we don't
do something about it.
Eric
|
|
|