Kind of Murphy
Of course it happens when I won’t be near the center: a power failure. This time it must have happened not that early since I could access before.
After work, I tried to access the router but the only thing that worked somewhat was the VPN connection – though unstable. Direct access would have worked if I would have used the proper protocol (https). But the server dis not respond to ICMP requests.
So I phoned home and learned about the power failure. I also learned there had been trouble with a secutity switch, and if it’s the one that serves the data center, it’s abvious it has been caused by the disks starting up. But at this moment it was clear that the disks were powered on. So I instructed my son to boot the system, and within a few minutes all was up and running again.
Well, almost. The public web couldn’t be accessed, and I found that startup didn’t configure it. But restarting the server solved it.
To determine the moment, I checked the logfile: it happened after 10:22 local time, the last line being:
%%%%%%%%%%% OPCOM 15-DEC-2010 10:22:57.95 %%%%%%%%%%%
Logfile time stamp
and DecEvent showed even more data:
**** V3.4 ********************* ENTRY 2362 ********************************
Logging OS 1. OpenVMS
System Architecture 2. Alpha
OS version V8.3
Event sequence number 29388.
Timestamp of occurrence 15-DEC-2010 10:57:40
Time since reboot 23 Day(s) 22:39:34
Host name DIANA
System Model AlphaServer DS10 617 MHz
Entry Type 38. Time Stamp Entry
SWI Minor class 7. Timestamp
I would expect to get even more precise infromation bu examining the SYSLOGD log – but found it hadn’t started since the last update – nor now. This is something to be looked at; manually starting the server using logger didn’t work …
I have set Syslog on port 517 in stead of the default 514. This port is also used by RSH – by the setting of HP OpenVMS CTPIP$CONFIG – but for TCP instead of UDP. But it might have caused a problem. However, the logger did now start properly – it may have started anyway but it has been overlooked. But no matter what, it seems that the boot after updates didn’t start SYSLOG.
Well, it now runs.