Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uwm.edu!lll-winken!ncis.tis.llnl.gov!blackbird!udecc!turner From: turner@udecc.engr.udayton.edu (Bob Turner) Newsgroups: comp.unix.i386 Subject: Re: Archive Tapes Message-ID: <1990May3.182056.9449@udecc.engr.udayton.edu> Date: 3 May 90 18:20:56 GMT References: <29490@cup.portal.com> <1990May2.113532.26951@msuinfo.cl.msu.edu> Reply-To: turner@udecc.engr.udayton.edu (Bob Turner) Organization: Univ. of Dayton, School of Engineering Lines: 45 In article <1990May2.113532.26951@msuinfo.cl.msu.edu> jhl@frith.uucp (John Lawitzke) writes: >From article <29490@cup.portal.com>, by Dante_A_Nicolello@cup.portal.com: >> Problem two: >> When the tape reaches the end of the cartridge, not the end of the >> current track, unix produces a "write() error" instead of prompting >> me to switch tapes. > >tar and cpio have no idea how long or how dense your tape is. They >assume you will not attempt to write more to tape than it can hold. If >you use the backup program you can specifiy density and length so the >program knows when it needs to prompt for a new tape. > >One solution is find the source to pdtar and hack it to prompt for a new >tape at the necessary point. > On a consulting job I did in March we had the same problem. Especially with cpio. We were using a 499 and a 60MB Fastape. Tar is an easy. It does not recognize the end of tape in any situation. But all is not lost. It keeps count of the number blocks written so far. And compares it the number of blocks available that was specified on the command line or in the file /etc/defaults/tar (or something like that). If you want specifics I will be happy to post. Cpio is a different beast. Cpio according to the manual :-), is supposed to detect the end of the device and prompt for a new filename. Usually you key in the tape again after you change the cartridge. Unfortunately, it actually works as you describe. The problem seems to lie within the device driver that Archive supplys. The device driver when it hits the end of the tape returns a write error #5 (or something approximating it) which is the general I/O Error rather than no more space on device (Error number 27). So that cpio knows to prompt. The unfortunate part is that the backup script also uses cpio. Realize that there is obvoiusly no problems unless you need to span a tape. If you need more help, mail me a message. Bob -- ==================================================================== Bob Turner Network Manager, School of Engineering 513-229-3171 turner@udecc.engr.udayton.edu Univ. of Dayton, Engineering Computing Center-KL211, Dayton OH 45469