Slashdot doesn't like me. Or my ISP. Or something. It timeouts on me. So I went to google and trawled the VAX story out from elseweb.
Reminds me of something I did in 198n.
Now, there's this particular make of air surveillance resources. It runs on a cluster of PDP-11's. The sites are tightly networked, using a proprietary networking system that ties the sites really tight together.
Now, I was used to doing things on a PDP-11 running RSX -- on a normal subscriber terminal. Some of the utility tools botched the terminal settings sometimes, and the simplest way to recover was hit the reset key on the terminal. Now, for this particular experiment I was not at the central site on a subscriber terminal. I was actually at the surveillance resource, tapping away at the console terminal. Logged in at operator privileges.
The utility did its usual weirdness, and the terminal was all unreadable.
As I had conditioned myself to do, I hit the terminal reset key.
Except that on those PDP's, when you press the reset key on the console terminal, you are actually resetting the computer itself.
Did I mention that the PDP's at each site are clustered? All round me I heard the bootup sequence of the computers, and the frantic thudding of the antenna control system hydraulics hitting failsafe positions. Resetting a single computer caused every computer in the cluster to halt.
Did I mention that the different sites are tightly networked? Loss of heartbeat from one site made several adjacent sites switch to recovery mode, which severely degrades the quality of output, while ensuring the existence of output.
It took less than an hour and we had the systems up and running as if nothing had happened. I've later checked discreetly, and all archive tapes have long since been discarded. No evidence exists that we had a gaping hole in our air surveillance capacity that summer morning.
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
Reminds me of something I did in 198n.
Now, there's this particular make of air surveillance resources. It runs on a cluster of PDP-11's. The sites are tightly networked, using a proprietary networking system that ties the sites really tight together.
Now, I was used to doing things on a PDP-11 running RSX -- on a normal subscriber terminal. Some of the utility tools botched the terminal settings sometimes, and the simplest way to recover was hit the reset key on the terminal. Now, for this particular experiment I was not at the central site on a subscriber terminal. I was actually at the surveillance resource, tapping away at the console terminal. Logged in at operator privileges.
The utility did its usual weirdness, and the terminal was all unreadable.
As I had conditioned myself to do, I hit the terminal reset key.
Except that on those PDP's, when you press the reset key on the console terminal, you are actually resetting the computer itself.
Did I mention that the PDP's at each site are clustered? All round me I heard the bootup sequence of the computers, and the frantic thudding of the antenna control system hydraulics hitting failsafe positions. Resetting a single computer caused every computer in the cluster to halt.
Did I mention that the different sites are tightly networked? Loss of heartbeat from one site made several adjacent sites switch to recovery mode, which severely degrades the quality of output, while ensuring the existence of output.
It took less than an hour and we had the systems up and running as if nothing had happened. I've later checked discreetly, and all archive tapes have long since been discarded. No evidence exists that we had a gaping hole in our air surveillance capacity that summer morning.