fouts@ames-nas-gw.arpa (10/04/84)
From: Martin Fouts <fouts@ames-nas-gw.arpa> Count me in. ----------
dave@RIACS.ARPA (10/04/84)
From: "David L. Gehrt" <dave@RIACS.ARPA> 1... dave ----------
dave@RIACS.ARPA (10/10/84)
From: "David L. Gehrt" <dave@RIACS.ARPA> I am about ready to make a limited distribution of a new version of a bad block forwarding driver for the UDA 50. This driver is a serious expansion of the dirver which was written by qtlon!tony. It does what I believe to be a complete job of implementing host initiated dynamic bad block replacement. We have been running it full time here for a few days with no particular problems after checking all but the most bizarre paths throught the code (for example, untested is the code for a completely full RCT, implying no more replacement blocks). I have not yet incorporated support for drive types other that the ra81, but I hope to have that installed and wrung out by the weekend. I am also doing a little document covering the methodology of replacement, and I have a program which dumps the replacement and caching tables to allow a little watch-dogging of the process. The purpose of this is to solicit the assistance of a few sites for beta testing. I believe the driver to be relatively solid, and therefore its use to be low risk. It does not appear to scramble file systems, and it hasn't crashed our production system but once, and that was my fault in making a sloppy fix to a performance problem. I would prefer to distribute the code to sites with access to the UDA 50 Programmers Kit, but such access is not an absolute necessity. If you are interested, please let me know. I hope to be ready by the middle of next week. I will be happy to answer questions from prospective beta test sites, of course. Oh, since the distribution will be well in excess of 100K bytes, I'd prefer to deal with arpa sites as beta test sites. Rather than have to split up. dave ----------