Description of problem: with F10 beta, when trying kickstart, my machine will sit indefinitely (longest wait was a full night) at the "waiting for hardware to initialize..." step. But if I press enter anaconda dies with the following traceback loader received SIGSEGV! Backtrace: /sbin/loader(loaderSegvHandler+0xa0)[0x804fd20] [0x132400] /lib/libnewt.so.0.52(newtGetKey+0x50)[0x136e70] /lib/libnewt.so.0.52(newtFormRun+0x44e)[0x139c9e] /lib/libnewt.so.0.52(newtRunForm+0x29)[0x139ed9] /sbin/loader(configureTCPIP+0x769)[0x8063e29] /sbin/loader(readNetConfig+0x7f)[0x8065ddf] /sbin/loader(kickstartNetworkUp+0x63)[0x8065fb3] /sbin/loader(getFileFromUrl+0xa8)[0x8060ee8] /sbin/loader(kickstartFromUrl+0x20)[0x8061470] /sbin/loader(getKickstartFile+0xc1)[0x8059801] /sbin/loader(main+0x2327)[0x80529b7] /lib/libc.so.6(__libc_start_main+0xe5)[0x3b16e5] /sbin/loader[0x804d981] install exited abnormally [1/1] Version-Release number of selected component (if applicable): Fedora 10 beta How reproducible: always Steps to Reproduce: 1. have a working PXE entry (F9, F8, RHEl5, RHEL4, etc) 2. modify it for F10 beta kickstart testing 3. try to kickstart a machine Actual results: failure to kickstart Expected results: successful kickstart Additional info: quite a few notes: 1) my menu.c32 entry is as follows, please do double check my sanity, I've looked over it a dozen times but may have missed a typo label f10betai386ks MENU LABEL Fedora 10 Beta i386 kickstart install TEXT HELP This entry will erase all your HDs Only use if you want to install a system for testing ENDTEXT IPAPPEND 2 kernel f10betai386/vmlinuz append initrd=f10betai386/initrd.img ks=http://192.10.10.250/kickstart/pcfe/F10-beta-minimaltesting-i386-ks.cfg ksdevice=bootif console=ttyS0,115200 (last 3 are obviously one line) 2) an interactive install on these boxes works just fine (still have bug 377921 but install itself in interactive mode does work) 3) This hardware is a bit different from run-off-the-mill hardware in the following points - there is no VGA hardware on the blade - all my HDs are attached through QLA2312 controllers (2 controllers per system I have tried this with one path up and both paths up)
Created attachment 319879 [details] log from the console server when this happens console server will insert timestamps but it does log everything nicely (including control chars)
Fixed in anaconda-11.4.1.49-1, please test rawhide when that version appears in the tree.
confirm this is gone in anaconda 11.4.1.50 was battling with what I filed today as Bug 467753, so did not get to test with .49 In any case, this can be CLOSED RAWHIDE I presume.