Just did 38.5 MB complete in 0:48, 798 kB/s, 64000 byte blocks. Kent Snook wrote: > > Pricop Ioan wrote: > > > > --- Kent Snook <kent.snook@echostar.com> wrote: > > > We encountered similiar problems. Implemented the > > > following. > > > > > > First results in an automatic reboot upon tftp > > > timeouts, instead of > > > waiting for a key press ( we regularly reboot these > > > hosts remotely ). > > > Second reduces the tftp timeout to get hosts up and > > > running much faster. > > > ( We use ramdisks for all of our diskless hosts, so > > > images can run up to > > > 6-10 mb in size, at 5 seconds per packet tftp > > > timeout, could take 15 > > > minutes or more to boot. Now less than a minute.) > > > > > .......... > > > > > > -- > > > Kent > > > > > > > Dear Kent, > > > > Maybe you can help me with my problem. I am using > > EtherBoot to experiment with a diskless client. After > > a few hoops, I made it working from a NFS-root > > filesystem. Now I am experimenting with RAM-disks. I > > tried with a 9 MB fs and it seems to work. Then, I > > tried with a 96 MB fs. Of cause, I modified the > > /usr/src/linux-2.2.5/drivers/block/rd.c to be able to > > create ramdisks of 128 MB (have 196 MB on the client) > > and recompiled the kernel. I tested these, by creating > > a 128 MB RAMdisk and populating it with the fs I > > downsized at 96MB. It worked with a "chroot", > > exception the X part (only fvwm and Netscape). I > > obtained a 50 MB "rdImage.gz". But when I tried to > > make a NBI and boot it, all is well until it's > > tftp-ing the NBI. Than the rotator keeps spinning and > > spinning forever - i think, cause after 2.5 hours I > > stopped the process. Anyway, with "tcpdump" I > > estimated the 3.8 MB composed immage loading time (fs > > of 9 MB gzipped) and calculated the 51 MB one's - it > > must have been 1 min and 47 secs. What can be wrong? > > Can anyone tip me on some sites where I can find > > specific documentation on NBI use of ramdisks? > > > > Thank you for at least reading this mail, > > > > ===== > > Ioan (John) NIKY Pricop > > Bucharest - ROMANIA > > pniky@yahoo.com > > > > A 96MB image could take many hours to tftp to the client. Before I > changed the TFTP_TIMEOUT value, a 10MB image could take 30 minutes on a > GOOD day. Turn on block counting with etherboot to get a visual of which > block you are on. If you require a filesystem as large as you are trying > to load, I would recommend using nfs instead of a RAMDISK. We took the > RAMDISK route to free our diskless hosts from dependencies on other > hosts on our net. > > We use make to run a Makefile which in turn executes shell commands and > calls shell scripts to create our images. Our largest image/filesystem > is around 10mB. I would recommend creating a filesystem that contains > only the binaries you actually require. Use ldd to determine the > libraries you need and eliminate the ones you don't need. Strip all > files (be carefull with any modules- don't use the --strip-all option > for these!!). > > Hope this helps. > > -- > Kent > =========================================================================== > This Mail was sent to netboot mailing list by: > Kent Snook <kent.snook@echostar.com> > To get help about this list, send a mail with 'help' as the only string in > it's body to majordomo@baghira.han.de. If you have problems with this list, > send a mail to netboot-owner@baghira.han.de. =========================================================================== This Mail was sent to netboot mailing list by: C J Considine <conx@connectnet.com> To get help about this list, send a mail with 'help' as the only string in it's body to majordomo@baghira.han.de. If you have problems with this list, send a mail to netboot-owner@baghira.han.de.
For requests or suggestions regarding this mailing list archive please write to netboot@gkminix.han.de.