Re: a few suggestions/comments on 3.0-RELEASE
Igor Roshchin wrote: I've encountered this problems while installing 3.0-RELEASE on a PII-350 with ASUS P2B-LS motherboard (i.e. there is Adaptec U2W adapter and Intel EtherExpress 100/10 on board) All HDDs are UW SCSI. 2. I didn't find any documentation in the man pages which describes the features of the new bootstrap - i.e. that it is capable of handling slice number. /boot.help doesn't have that reference either. (I mean the possibility of specifying 2:da(1,2,a)kernel where the second 2 is for the slice number) The only place I found where it was addressed was somebody's answer in -questions archive, and on the -stable list. The documentation is slowly being updated to reflect the features of the new boot code. There's nothing about specifying slices yet; though the preferred way to accomplish most boot stuff in 3.x is by way of the all-singing, all-dancing /boot/loader program rather than boot2. Also, it might be nice if the disklabel(8) man pages explicitely say that the bootstrap code can be installed on each slice. (am I wrong ?) You're right, the bootblocks can be installed on each slice, and this also needs documenting. -- Robert Nordier To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: a few suggestions/comments on 3.0-RELEASE
Also, it might be nice if the disklabel(8) man pages explicitely say that the bootstrap code can be installed on each slice. (am I wrong ?) You're right, the bootblocks can be installed on each slice, and this also needs documenting. This doesn't need documenting. disklabel(8) knows nothing of slices. It installs labels and bootblocks on disk-like devices. Slices are just one type of disk-like device. All regular files larger than 1K and many devices are disk-like, thanks to device independence. For the bootstrap code to actual work, more is required. E.g., slices on vn disks are not bootable because the boot loader doesn't support vn disks. disklabel(8) knows nothing of this limitation. Bruce To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: a few suggestions/comments on 3.0-RELEASE
Bruce Evans wrote: Also, it might be nice if the disklabel(8) man pages explicitely say that the bootstrap code can be installed on each slice. (am I wrong ?) You're right, the bootblocks can be installed on each slice, and this also needs documenting. This doesn't need documenting. disklabel(8) knows nothing of slices. It installs labels and bootblocks on disk-like devices. Slices are just one type of disk-like device. All regular files larger than 1K and many devices are disk-like, thanks to device independence. I think one does well to avoid No! No! No! reponses to useful feedback. The original suggestion highlights aspects of the new bootblocks that are incompletely documented and potentially confusing. My reply was intentionally phrased to acknowledge this while omitting all reference to the disklabel man page. This does need documenting. disklabel(8) is merely an installation tool for bootstrap stages 1 2, as is install(1) for bootstrap stage 3. Of course one does not document programs on the man pages of their installation tools. However this does not mean that one does not document them (or their installation procedures) at all. -- Robert Nordier To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
a few suggestions/comments on 3.0-RELEASE
Hello! First of all I appologize if any of these problems have been addressed on the -current list, I am not subscribed, and at the moment don't have much time to carefully browse through the archive of the -current. However, after a quick glance I didn't find it reported. All this is related to the 3.0-RELEASE. I've encountered this problems while installing 3.0-RELEASE on a PII-350 with ASUS P2B-LS motherboard (i.e. there is Adaptec U2W adapter and Intel EtherExpress 100/10 on board) All HDDs are UW SCSI. 1. Having two /stand/sysinstall working in parallel in two different virtual consoles (after the system is brought up after the initial installation), makes the system to panic and reboot. This seems to happen if, and at the moment when both sysinstall's are writing something intensively, say installing packages in one and installing additional options for the distribution (catpages, manpages, sources..) - both from the ftp. The errors appear on the screen so fast that I couldn't catch what they were, but they seemed to be related to the disk i/o. Everytime (I tried 3-4 times) it was the same error. None is written to /var/log/messages. 2. I didn't find any documentation in the man pages which describes the features of the new bootstrap - i.e. that it is capable of handling slice number. /boot.help doesn't have that reference either. (I mean the possibility of specifying 2:da(1,2,a)kernel where the second 2 is for the slice number) The only place I found where it was addressed was somebody's answer in -questions archive, and on the -stable list. Also, it might be nice if the disklabel(8) man pages explicitely say that the bootstrap code can be installed on each slice. (am I wrong ?) 3. if I start /stand/sysinstall after bringing the computer up for the first time, and try to add and configure the network interface (fxp0 in my case), it prompts me if I want to bring it up. If I say yes, it was not complaining about anything, but in some cases (may be even all - I didn't test extensively) the routing to the will not come up (according to the netstat), so it'd look like (approximately, recreated from my memory): netstat -i Name Mtu Network AddressIpkts IerrsOpkts Oerrs Coll fxp0 1500 Link 00.00.c0.3d.41.c6 0 00 0 0 lp0* 1500 Link 0 00 0 0 tun0* 1500 Link 0 00 0 0 lo0 16384 Link 0 00 0 0 lo0 16384 127 localhost 0 00 0 0 Just doing the configuration of the interface, and than rebooting does the job alright. (My only doubt is that I don't remember whether there were any leftovers of attempts of configuring the fxp0 interface manually just before starting the sysinstall. In any case I think sysinstall should be able to bring the interface up properly.) I didn't experience any problem with bringing the interface up from the installation floppy. 4. (it's probably relevant to -ports list ?) make for the ssh (not ssh2) complains about the absence of bsd.port.post.mk and bsd.port.pre.mk Copying it from the -stable brunch seems to fix the problem. (It seems that I've seen somebody complaining about this problem, but couldn't find it anywhere, and it is not reflected in the ERRATA http://www.freebsd.org/releases/3.0R/errata.html ) If you want to reply or to ask about some details I missed, please include my address in the reciepients, because I am not subscribed to -current mailing list. Regards, Igor To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message