After my text install failed, I tried VNC. There were some strange black areas during the package selection screen (http://david.woodhou.se/vnc3.jpeg) and then after formatting the file systems and starting the install, the VNC window just went black (http://david.woodhou.se/vnc4.jpeg).
The VNC server is now running. Please connect to uranus.cambridge.redhat.com:1 to begin the install... Starting graphical installation... Press <enter> for a shell Unable to read wel help text Unable to read autopart help text Unable to read netconf help text Unable to read timezone help text Unable to read accts help text Unable to read aboutinstall help text Unable to read installing help text sh-3.00# ps axf PID TTY STAT TIME COMMAND 1 ? S 0:00 /init vnc 2 ? S 0:00 [migration/0] 3 ? SN 0:00 [ksoftirqd/0] 4 ? S 0:00 [watchdog/0] 5 ? S 0:00 [migration/1] 6 ? SN 0:00 [ksoftirqd/1] 7 ? S 0:00 [watchdog/1] 8 ? S 0:00 [migration/2] 9 ? SN 0:00 [ksoftirqd/2] 10 ? S 0:00 [watchdog/2] 11 ? S 0:00 [migration/3] 12 ? SN 0:00 [ksoftirqd/3] 13 ? S 0:00 [watchdog/3] 14 ? S 0:00 [migration/4] 15 ? SN 0:00 [ksoftirqd/4] 16 ? S 0:00 [watchdog/4] 17 ? S 0:00 [migration/5] 18 ? SN 0:00 [ksoftirqd/5] 19 ? S 0:00 [watchdog/5] 20 ? S 0:00 [migration/6] 21 ? SN 0:00 [ksoftirqd/6] 22 ? S 0:00 [watchdog/6] 23 ? S 0:00 [migration/7] 24 ? SN 0:00 [ksoftirqd/7] 25 ? S 0:00 [watchdog/7] 26 ? S< 0:00 [events/0] 27 ? S< 0:00 [events/1] 28 ? S< 0:00 [events/2] 29 ? S< 0:00 [events/3] 30 ? S< 0:00 [events/4] 31 ? S< 0:00 [events/5] 32 ? S< 0:00 [events/6] 33 ? S< 0:00 [events/7] 34 ? S< 0:00 [khelper] 35 ? S< 0:00 [kthread] 42 ? S< 0:00 \_ [kblockd/0] 43 ? S< 0:00 \_ [kblockd/1] 44 ? S< 0:00 \_ [kblockd/2] 45 ? S< 0:00 \_ [kblockd/3] 46 ? S< 0:00 \_ [kblockd/4] 47 ? S< 0:00 \_ [kblockd/5] 48 ? S< 0:00 \_ [kblockd/6] 49 ? S< 0:00 \_ [kblockd/7] 53 ? S< 0:00 \_ [khubd] 112 ? S 0:00 \_ [pdflush] 113 ? S 0:00 \_ [pdflush] 115 ? S< 0:00 \_ [aio/0] 116 ? S< 0:00 \_ [aio/1] 117 ? S< 0:00 \_ [aio/2] 118 ? S< 0:00 \_ [aio/3] 119 ? S< 0:00 \_ [aio/4] 120 ? S< 0:00 \_ [aio/5] 121 ? S< 0:00 \_ [aio/6] 122 ? S< 0:00 \_ [aio/7] 193 ? S< 0:00 \_ [khvcd] 205 ? S< 0:00 \_ [kseriod] 618 ? S< 0:00 \_ [scsi_eh_0] 623 ? S< 0:00 \_ [scsi_eh_1] 637 ? S< 0:00 \_ [rpciod/0] 638 ? S< 0:00 \_ [rpciod/1] 639 ? S< 0:00 \_ [rpciod/2] 640 ? S< 0:00 \_ [rpciod/3] 641 ? S< 0:00 \_ [rpciod/4] 642 ? S< 0:00 \_ [rpciod/5] 643 ? S< 0:00 \_ [rpciod/6] 644 ? S< 0:00 \_ [rpciod/7] 820 ? S< 0:00 \_ [xfslogd/0] 821 ? S< 0:00 \_ [xfslogd/1] 822 ? S< 0:00 \_ [xfslogd/2] 823 ? S< 0:00 \_ [xfslogd/3] 824 ? S< 0:00 \_ [xfslogd/4] 825 ? S< 0:00 \_ [xfslogd/5] 826 ? S< 0:00 \_ [xfslogd/6] 827 ? S< 0:00 \_ [xfslogd/7] 828 ? S< 0:00 \_ [xfsdatad/0] 829 ? S< 0:00 \_ [xfsdatad/1] 830 ? S< 0:00 \_ [xfsdatad/2] 831 ? S< 0:00 \_ [xfsdatad/3] 832 ? S< 0:00 \_ [xfsdatad/4] 833 ? S< 0:00 \_ [xfsdatad/5] 834 ? S< 0:00 \_ [xfsdatad/6] 835 ? S< 0:00 \_ [xfsdatad/7] 836 ? S< 0:00 \_ [xfsbufd] 871 ? S< 0:00 \_ [kmirrord] 1044 ? S< 0:00 \_ [kauditd] 74 ? S 0:00 [rtasd] 114 ? S 0:00 [kswapd0] 427 ? S 0:00 /init vnc 428 ? S 0:03 /sbin/loader --virtpconsole /dev/hvc0 675 tty2 Ss+ 0:00 \_ -/bin/sh 885 ? S 0:02 \_ /usr/bin/python /usr/bin/anaconda -m nfs://mnt/so 904 ? S 0:02 \_ //usr/bin/Xvnc :1 -nevershared -depth 16 -geo 905 ? S 0:00 \_ //usr/bin/mini-wm --display :1 908 ? Z 4:05 \_ [anaconda] <defunct> 1050 ? S 0:00 \_ /bin/sh 1051 ? R 0:00 \_ ps axf 645 ? S< 0:00 [loop0] 788 ? S 0:00 [jfsIO] 789 ? S 0:00 [jfsCommit] 790 ? S 0:00 [jfsCommit] 791 ? S 0:00 [jfsCommit] 792 ? S 0:00 [jfsCommit] 793 ? S 0:00 [jfsCommit] 794 ? S 0:00 [jfsCommit] 795 ? S 0:00 [jfsCommit] 796 ? S 0:00 [jfsCommit] 797 ? S 0:00 [jfsSync] 1038 ? S 0:00 [kjournald] 1039 ? S 0:00 [kjournald] 1047 ? S 0:00 /usr/bin/python /usr/bin/anaconda --syslogd /mnt/sysi sh-3.00# cat /mnt/sysimage/root/install.log Installing libgcc-4.1.0-0.6.ppc64. Installing libgcc-4.1.0-0.6.ppc. Installing setup-2.5.47-1.1.noarch. warning: setup-2.5.47-1.1: Header V3 DSA signature: NOKEY, key ID 30c9ecf8 Installing filesystem-2.3.7-1.ppc. Installing fedora-logos-1.1.34-1.noarch. Installing gnome-mime-data-2.4.2-1.ppc. warning: gnome-mime-data-2.4.2-1: Header V3 DSA signature: NOKEY, key ID 4f2a6fd2 Installing basesystem-8.0-5.noarch. Installing mailcap-2.1.19-1.noarch. Installing termcap-5.4-7.noarch. Installing cracklib-dicts-2.8.6-1.ppc. Installing xorg-x11-xkbdata-0.99.1-2.noarch. Installing rmt-0.4b40-5.ppc. Installing slib-3a1-5.noarch. Installing desktop-backgrounds-basic-2.0-30.1.noarch. Installing iso-codes-0.47-1.noarch. Installing bsf-2.3.0-6jpp_2fc.noarch. Installing tzdata-2005n-2.noarch. Installing glibc-common-2.3.90-18.ppc.
stracing the installer didn't enlighten me much. It deadlocks if I do that, when installing the first package... clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0xf7ffa048) = 1061 rt_sigprocmask(SIG_SETMASK, NULL, [CHLD], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_SETMASK, NULL, [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [CHLD], NULL, 8) = 0 close(72) = 0 close(73) = 0 gettimeofday({1134264961, 31036}, NULL) = 0 rt_sigprocmask(SIG_SETMASK, NULL, [CHLD], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 --- SIGCHLD (Child exited) @ 0 (0) --- waitpid(0, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], WNOHANG) = 1061 waitpid(0, 0xffef6b9c, WNOHANG) = 0 rt_sigreturn(0) = 0 futex(0x12c78504, FUTEX_WAIT, 1, NULL
Upon reproducing with 'debug' I see it get as far as installing glibc-common again before it stops, although this time the screen isn't black and anaconda still appears to be running. There's interesting stuff in /mnt/sysimage/root/anaconda.log now though... Installing libgcc-4.1.0-0.6.ppc64. Installing libgcc-4.1.0-0.6.ppc. Installing setup-2.5.47-1.1.noarch. warning: setup-2.5.47-1.1: Header V3 DSA signature: NOKEY, key ID 30c9ecf8 Installing filesystem-2.3.7-1.ppc. Installing fedora-logos-1.1.34-1.noarch. Installing gnome-mime-data-2.4.2-1.ppc. warning: gnome-mime-data-2.4.2-1: Header V3 DSA signature: NOKEY, key ID 4f2a6fd2 Installing basesystem-8.0-5.noarch. Installing mailcap-2.1.19-1.noarch. Installing termcap-5.4-7.noarch. Installing cracklib-dicts-2.8.6-1.ppc. Installing xorg-x11-xkbdata-0.99.1-2.noarch. Installing rmt-0.4b40-5.ppc. Installing slib-3a1-5.noarch. Installing desktop-backgrounds-basic-2.0-30.1.noarch. Installing iso-codes-0.47-1.noarch. Installing bsf-2.3.0-6jpp_2fc.noarch. Installing tzdata-2005n-2.noarch. Installing glibc-common-2.3.90-18.ppc. warning: user H does not exist - using root warning: user \uffff\uffff does not exist - using root warning: user an1/GET.1.gz does not exist - using root warning: user gz does not exist - using root warning: user init/Xclients.d does not exist - using root warning: user .0 does not exist - using root warning: user \uffff$T does not exist - using root warning: user /suexec.8.gz does not exist - using root warning: user ec.8.gz does not exist - using root warning: user gz does not exist - using root warning: user t/ does not exist - using root warning: user \uffff\uffff\uffff does not exist - using root warning: user r/share/man/man8/httpd.8.gz does not exist - using root warning: user n/man8/httpd.8.gz does not exist - using root ...etc.... If I now attach with strace, I see it's just doing this... futex(0x1285bf70, FUTEX_WAIT, 2, NULL) = -1 EAGAIN (Resource temporarily unavailable) futex(0x1285bf70, FUTEX_WAIT, 2, NULL) = -1 EAGAIN (Resource temporarily unavailable) futex(0x1285bf70, FUTEX_WAIT, 2, NULL) = -1 EAGAIN (Resource temporarily unavailable) The package appears to be fine... $ md5sum glibc-common-2.3.90-18.ppc.rpm ; rpm --checksig glibc-common-2.3.90-18.ppc.rpm 00ef660bedf55603d9d396311add9fd1 glibc-common-2.3.90-18.ppc.rpm glibc-common-2.3.90-18.ppc.rpm: (SHA1) DSA sha1 md5 (GPG) NOT OK (MISSING KEYS: GPG#30c9ecf8)
The black screen is all Xvnc -- there's really not much we do which can influence that one way or another :)
The file system panic is fixed by reverting the recent changes to the service partition, and the install completed successfully (modulo selinux bugs qv)
David: does the black-screen-from-vnc-install problem still happen after reverting those changes?
The final all-black screen (vnc4.jpeg), which I think happened because anaconda had crashed or aborted, no longer happens. However, I do still see the partial areas of black, such as in http://david.woodhou.se/vnc3.jpeg and http://david.woodhou.se/vnc-part2.jpeg. Note that the black region in vnc-part2.jpeg goes only black when the top-most dialog box pops up; until then it was fine (cf. vnc-part1.jpeg).
From the look of the screenshot you are using a limited colour palette; perhaps enabling full colour (F8->Options) will give better results? Anyway, it seems as though VNC isn't really the issue here.
Reassigning.
Full colour doesn't help -- it makes the non-black bits slightly prettier, but the black parts remain black.
Is this still occurring? I haven't seen it at all (and have done a lot of vnc installs for Xen)
I had a report the other day that it was still happening -- will test when I get home next week.
Black backing for vnc on ppc seen again today on a install on a ppc box with VNC - if you need access to a ppc box to test please let me know, and one can be arranged.
I can confirm this with 4.92: http://lisas.de/~adrian/vnc1.png http://lisas.de/~adrian/vnc2.png
And with the final release of FC5.
*** Bug 197677 has been marked as a duplicate of this bug. ***
We really need to fix this - Jitka Kudrnacova please let me know if you need access to a system to debug.
changed: What |Removed |Added ---------------------------------------------------------------------------- Summary|RH197677- fields in certain |RH175458- fields in certain |installation dialog windows |installation dialog windows |appear black |appear black ------- Additional Comments From chavez.com(prefers email via lnx1138.com) 2006-07-05 10:48 EDT ------- Cross-referencing over to RH BZ 175458 which looks to be a FC5 bug. I expect the same fix will end up in RHEL 5? So for my own benefit, had we opened both an FC5 and RHEL 5 bug reporting the same bug (without there being either already opened to dup to) would one have been dup'd to the other?
I have found out, that these black areas occur in all gtk applications, not only in anaconda. Non-gtk applications are not affected.
Why is this only FC6Target and not FC6Blocker?
Still in FC6 even now. This was reported before FC5. When is it going to get fixed?
*** Bug 174383 has been marked as a duplicate of this bug. ***
*** This bug has been marked as a duplicate of 192289 ***