dodgson@cs.wmich.edu (Harry Dodgson) (03/16/91)
I have upgraded the servers here to SunOS 4.1.1. I have crashed the tapehost twice so far trying to access a tape from a remote machine. This is the configuration each time: try 1. tapeHost Sun3/60, 16Meg RAM, 32Meg Swap, 1/4" tape drive client Sun4/280, 32Meg RAM, 80Meg Swap running extract_unbundled to install SUN Fortran 1.3.1 try 2. tapehost Sun4/470, 32Meg RAM, >64Meg Swap, 8mm tape client Sun3/60, 16Meg RAM, 32Meg Swap (same as above) running rrestore with "ifs" switches. Both times the tapehost paniced, crashed, and rebooted. Both tapehosts are acting as servers, the 4/470 is the primary NIS server and exports most of it's file structure - the 3/60 is secondary NIS server and exports one directory. All three units were recently upgraded to SunOS 4.1.1 and have only a few local customizations to /etc for our site. Both servers also run DNS if that matters. Our department currently does not receive the SUN bug fixes so I don't know if this has been documented yet. Harry Dodgson Jr. | Internet dodgson@sol.cs.wmich.edu -(35.132.4.2)- Western Michigan University | UUCP ...uunet!sharkey!wmichgw!wmu-cs!dodgson Computer Science Department | Voice (616) 387-5803 Kalamazoo, MI 49008 | Office 4420 Dunbar Hall
eggert@uunet.uu.net (Paul Eggert) (03/23/91)
dodgson@cs.wmich.edu (Harry Dodgson) writes that his remote SunOS 4.1.1 tapehosts crashed. We had a similar problem with remote dumping with a 4.1.1 Sun-3/60 8mm tapehost. At first the kernel paniced with a ``psig action'' message. Sun sent us a patch for bugid 1032534, but the dumps still fail with the following message: st1: tape synchronization lost psig: "rmt" signal 14 was masked, put back. Sun knows about this problem, but doesn't have a fix yet. As a stopgap we moved our tape drive to a 4.1.1 Sparcstation 1+, and we are crossing our fingers.