¿ªÔÆÌåÓý

PRO TIP (ooops!)


 

Pro Tip: Don't add a typewriter console to DMKRIO and then tell Hercules that device address is a card reader.

The combination sent a VM SIXPACK system from 1% host CPU to 100% host CPU, and internally had CP LOAD IND showing 79% CPU utilization without CPWATCH showing any users busy.

I was using the default DMKRIO for the SIXPACK systems and my default (included) configuration for for S/370 hosts. Hence the incompatibility.?

-ahd-


 

On Sat, Aug 17, 2024 at 05:16 PM, Drew Derbyshire wrote:
Pro Tip: Don't add a typewriter console to DMKRIO and then tell Hercules that device address is a card reader.
Does that mean it could give itself instructions independent of a human? ;-)
?
?... Mark S.


 

Hello!
Mark? That's a good one. Drew we all make mistakes. To really overdo
things we need to be the computers we write programs for. Strange, I
did expect a rather sharp piscine comment here.
-----
Gregg C Levine gregg.drwho8@...
"This signature fought the Time Wars, time and again."

On Sat, Aug 17, 2024 at 8:05?PM Mark A. Stevens via groups.io
<marXtevens@...> wrote:

On Sat, Aug 17, 2024 at 05:16 PM, Drew Derbyshire wrote:

Pro Tip: Don't add a typewriter console to DMKRIO and then tell Hercules that device address is a card reader.

Does that mean it could give itself instructions independent of a human? ;-)

... Mark S.


 

Gregg Levine wrote:

[...]
Strange, I did expect a rather sharp piscine comment here.
I would have, but I was taking a nap and missed my opportunity. :)

Besides, I've done much worse in my time. Very early in my career, I accidentally destroyed our sysres, and then after mounting our backup sysres, proceeded to accidentally destroy that one too! (Oops!)

It took us the entire night shift to completely rebuild it from scratch.

--
"Fish" (David B. Trout)
Software Development Laboratories

mail: fish@...


 

On Sat, Aug 17, 2024 at 10:28 PM, Fish Fish wrote:
Besides, I've done much worse in my time. Very early in my career, I accidentally destroyed our sysres, and then after mounting our backup sysres, proceeded to accidentally destroy that one too! (Oops!)

It took us the entire night shift to completely rebuild it from scratch.
No tape backup?
?
In ~ 1984, a friend related that at a previous job he likewise clobbered both sysres and its backup disk. He calmly told the operator to get the last night's backup. The operator told him that backups were taking too long, so the night shift had stopped running them. That is when the sweating started (and a similar rebuild from scratch was done).?

That story in turn reminds me that when I reorganized user packs with FDR, I did it after backups (which wrote two copies to different 9 track tapes) and than had the FDR reorg program write two more tapes. I wasn't gonna lose a pack because of laziness.
?
All this explains why at 2:39 AM every morning, my Hercules host dumps my emulated systems via tar. It does a differential backup nightly and a full 24 GB dump weekly. They go to a 1 TB Micro SD card, which is good for a ~ year of monthly backups. (It prunes the weekly backups after 90 days).? It's not trivial to get a single file off the backup, but I've done it.
?
-ahd-


 

Drew Derbyshire wrote:
wrote:

Besides, I've done much worse in my time. Very early in my
career, I accidentally destroyed our sysres, and then after
mounting our backup sysres, proceeded to accidentally destroy
that one too! (Oops!)

It took us the entire night shift to completely rebuild it
from scratch.
No tape backup?
It was so long ago, I can't recall why we didn't, no.

I do recall that we did do full volume stand-alone IPLable "Fast Copy?" backups of sysres (and made a Ditto copy of it afterwards, as well as a few other critical tapes too, which we kept hanging on a small tape rack in the machine room) which were then physically walked down the street to Bank by the Data Control department where had several rented Safety Deposit boxes, but I can't recall whether that procedure was in place at the time of my incident. I don't think it was because it would have made recovering a lot faster and easier. That procedure might have been begun *afterwards* BECAUSE of my incident. I'm thinking that at the time they felt just having a simple 3330 disk pack copy in the machine room that you could quickly swap out was good enough.

Hind sight is always 20/20, isn't it? :)


In ~ 1984, a friend related that at a previous job he likewise
clobbered both sysres and its backup disk. He calmly told the
operator to get the last night's backup. The operator told him
that backups were taking too long, so the night shift had stopped
running them. That is when the sweating started (and a similar
rebuild from scratch was done).
(Ouch!)


That story in turn reminds me that when I reorganized user packs
with FDR,
Yes! I think that's what we used too! (I think?)


I did it after backups (which wrote two copies to different 9 track
tapes)
I think we just made ditto copies?


and then had the FDR reorg program write two more tapes. I wasn't
gonna lose a pack because of laziness.

All this explains why at 2:39 AM every morning, my Hercules host
dumps my emulated systems via tar. It does a differential backup
nightly and a full 24 GB dump weekly.
I only do Fulls monthly. Daily differentials and monthly fulls. And before Microsoft dropped support for Windows 7, I'd do my fulls *before* applying my monthly patches.


They go to a 1 TB Micro SD card, which is good for a ~year of monthly
backups. (It prunes the weekly backups after 90 days).
I've got ... (me, checks) ... Yep! A year's worth of backups too.


It's not trivial to get a single file off the backup, but I've done
it.
What product do you use? I myself use a product called "Macrium Reflect" () and restoring individual files or an entire volume is a piece of cake.

It amazes me how many people don't bother backing up their system. Fools. Each and every one of them.

--
"Fish" (David B. Trout)
Software Development Laboratories

mail: fish@...