[2576] in SIPB-AFS-requests
ronald-ann outage at 1am, 11 December
daemon@ATHENA.MIT.EDU (mhpower@MIT.EDU)
Thu Dec 12 03:02:00 1996
From: mhpower@MIT.EDU
To: sipb-afsreq@MIT.EDU
Date: Wed, 11 Dec 96 01:47:57 EST
Earlier tonight there were some as-yet-not-fully-diagnosed problems
with ronald-ann, e.g.,
- it logged a message containing the text "was killed in xalloc:
no swap space"
- attempting to rlogin produced the error message "rcmd: bad
connection with remote host"
- attempting to bos exec the "ls" command gave the error message
"bos: failed to execute command (Unknown code a_ 0 (32512))"
- the bos exec was nevertheless logged normally by afs2sl
- fingering the machine resulted in a response of two blank lines,
rather than the machine type, load information, and the text
"No one logged on"
- it was possible to enter the username "root" at the console
prompt, but after 7 minutes there was still no password prompt
Despite this, no problems were noticed in its ability to serve files
via AFS. However, there was some speculation that the problems might
worsen and cause files and directories to become inaccessible,
although we were unable to agree on an adverb to describe how quickly
we thought this might happen. At about 1am, the command
/afs/sipb/service/@sys/bos shutdown -server ronald-ann -wait -cell sipb.mit.edu
was run. After that, the crash-dump instructions located on a nearby
piece of paper were followed. Once the partial dump succeeded,
control-C was typed and the machine and all of the disks were turned
off. The power strip was then moved to the active UPS that was labeled
with the username "kretch", and then the machine and disks were turned
back on. The AFS (R) 3.3 fsck program did not produce any Interesting
Messages. No files were reported lost in the SalvageLog file.
Matt
