Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 15316 - first "emerge -u system" fails with access denied errors
Summary: first "emerge -u system" fails with access denied errors
Status: RESOLVED FIXED
Alias: None
Product: Portage Development
Classification: Unclassified
Component: Unclassified (show other bugs)
Hardware: x86 Linux
: High major (vote)
Assignee: Nicholas Jones (RETIRED)
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2003-02-08 09:07 UTC by James Bowlin
Modified: 2011-10-30 22:21 UTC (History)
2 users (show)

See Also:
Package list:
Runtime testing required: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description James Bowlin 2003-02-08 09:07:55 UTC
Machine: Dell i8200 Laptop w/ Intel(R) Pentium(R) 4 Mobile CPU 1.60GHz

I'm trying to install Gentoo from a CD: stage3-pentium3-1.4_rc2.tbz2
Have installed okay on an ext2 partition yesterday.  
The "emerge -u system" failures occured when I was trying to
do the intial installation as per
http://www.gentoo.org/doc/en/gentoo-x86-install.xml
on a reiserfs partition w/ noatime, notail.

The access denied error messages were similar to those of
a previous bug report (13688) I hope I might be able to
provide more info to help you solve the problem.

The error messages started spewing while emerging xfree. Example:
make[5]: entering directory:
/var/tmp/portage/xfree-4.2.1-r2/work/xc/lib/font/X-TrueType
sh: line 1 /usr/bin/cc: no such file or directory

ACCESS DENIED chown: /var/cache/edb
ACCESS DENIED chown: /var/cahce/edb/dep
ACCESS DENIED open_wr: /var/cache/edb/mtimedb

Error in sys.exitfunc
traceback (most recent call last)
File "/usr/lib/python2.2/atexit.py line 20, in _run_exitfuncs
apply(func, targs, kargs)
File "/usr/lib/python2.2/site-packages/portage.py line 4236 in store
cPickle.dump(mtimedb, open(mymcn, "w")
IOError [errno 13] Permission Denied: '/var/cache/edb/mtimedb'

Before these failures I noticed a couple of things:

1) At the end of emerging gcc, a sanity check failed for
   g++ (I think, perhaps a different prog.) and emerge reverted
   back to the saved copy.

2) My CFLAGS from the /etc/make.conf file were not being used.
CLFAGS="-O3 -march=pentium4 -fforce-addr -fomit-frame-pointer -funroll-loops 
-malign-functions=4 -pipe"

3) Some of the .tbz2 files on the CD in /gentoo/packages/All
   were corrupt.  Example:

   # tar jtf /mnt/cdrom/gentoo/packages/All/kde-3.0.4.tbz2

   bzip2: (stdin): trailing garbage after EOF ignored

   I know these were real errors because the lists (that I
   tried to install) that had broken .tbz2 files did not
   install properly yesterday on the ext2 partition.  I did
   not notice an warning or failure message but it may have
   flown by.

4) Since I had just done an install yesterday on a different
   partition on the same machine before the emerge -u system
   I copied the contents of /usr/portage/distfiles/* from the
   old partition to the new one.

5) Before the problems started happening I had Ctrl-C'ed out
   of the emerge when it was doing that 5 4 3 2 1 countdown.
   I then restarted the emerge (looking to see if my CFLAGS
   would now be used).

6) After the ACCESS DENIED messages I Ctrl-C'ed out of emerge
   in the middle of a long string of error messages.  I then
   tried to umount everything I had mounted so I could reboot
   gracefully.  I was unable to umount the reiserfs.

   # umount /mnt/gentoo
   umount: device is busy

   # umount -f /mnt/gentoo
   umount2: Device or resource is busy
   umount: /dev/hda8 not mounted        # the reiserfs partition
   umount: /dev/gentoo: Illegal Seek

   I had exited out the chroot and had removed mounts that
   had been mounted on top of /dev/gentoo such as /dev/gentoo/boot
   and /dev/gentoo/proc.   I was able to "mount -o remount -o ro ..."
   so that I could run reiserfsck which said that there was nothing
   wrong.  I still couldn't umount.  I rebooted (in to the ext2 Gentoo
   system) and had no trouble with the reiserfs.

7) I was using all 4 virtual consoles at the time of the trouble.
   I was running things like "emerge -up system" while the 
   "emerge -u system" was running.

Reproducible: Didn't try
Steps to Reproduce:
1. emerge sync
2. emerge -up system
3. emerge -u system

More details are in teh Details section.

I have not tried to reproduce the error.  So many things seemed
to go wrong all of a sudden I am not sure what direction to take.
Actual Results:  
See Details

Expected Results:  
I expected it to emerge all of my "system" packages so I could then
download and build a kernel.

I can't run "emerge info" without rebooting and destroying all of the information
I have already entered.  I will submit this now and see if I can reboot into
the new system and run "emerge info".

re: Severity
For me it seems to be Critical or a Blocker.  But since it may be
entirely due to pilot error, for you it may be more like normal.
I compromised with Major.
Comment 1 Seemant Kulleen (RETIRED) gentoo-dev 2003-02-08 22:33:42 UTC
after you emerge sync, please emerge portage, then emerge -u world
Comment 2 James Bowlin 2003-02-08 23:13:04 UTC
I just tried:

emerge sync
emerge portage
emerge -u world

after starting over from scratch again.

Same problem.  Nothing changes AFAIK.  Same ACCESS DENIED errors.
Comment 3 Martin Holzer (RETIRED) gentoo-dev 2003-02-09 09:40:40 UTC
could you paste
#emerge info
Comment 4 James Bowlin 2003-02-09 14:03:30 UTC
# emerge info 
Portage 2.0.44 (default-x86-1.4, gcc-sh: line 1: /usr/bin/gcc: No such file or directory, 
glibc-2.3.1-r2) 
================================================================= 
System uname: 2.4.19-gentoo-r10 i686 Intel(R) Pentium(R) 4 Mobile CPU 1.60GHz 
USE="oss avi cups encode gpm imlib jpeg libg++ mikmod mmx motif nls oggvorbis qtmt sdl 
svga xmms xv zlib gdbm slang arts atlas apm berkdb cdr crypt dvd gd ggz gif gtk java jikes 
libwww mozilla mpeg ncurses opengl pam pcmcia pdflib perl plotutils python png pnp 
quicktime qt readline spell sse ssl tcpd truetype tcltk usb X x86 xml2 kde tiff -gnome -3dnow" 
ARCH="x86" 
COMPILER="gcc3" 
CHOST="i686-pc-linux-gnu" 
CFLAGS="-O3 -march=pentium4 -fforce-addr -fomit-frame-pointer -funroll-loops 
-falign-functions=4 -pipe" 
CXXFLAGS="-O3 -march=pentium4 -fforce-addr -fomit-frame-pointer -funroll-loops 
-falign-functions=4 -pipe" 
ACCEPT_KEYWORDS="x86" 
CONFIG_PROTECT="/etc /var/qmail/control /usr/kde/2/share/config /usr/kde/3.1/share/config 
/usr/kde/3/share/config /usr/share/config" 
CONFIG_PROTECT_MASK="/etc/gconf /etc/env.d" 
MAKEOPTS="-j2" 
JDK_HOME="/opt/blackdown-jdk-1.3.1" 
JAVA_HOME="/opt/blackdown-jdk-1.3.1" 
AUTOCLEAN="yes" 
SYNC="rsync://rsync.gentoo.org/gentoo-portage" 
GENTOO_MIRRORS="http://gentoo.oregonstate.edu/                  
http://csociety-ftp.ecn.purdue.edu/pub/gentoo/                 
http://distro.ibiblio.org/pub/Linux/distributions/gentoo" 
 
BTW: I had the same problems running with the default CFLAGS/CXXFLAGS 
Comment 5 Martin Holzer (RETIRED) gentoo-dev 2003-02-09 14:33:26 UTC
seems to be the gcc issue
Comment 6 Martin Holzer (RETIRED) gentoo-dev 2003-02-09 15:06:23 UTC
sorry but http://cvs.gentoo.org/~seemant/gcc.html is no more available

seemant: where's this document available ?
Comment 7 James Bowlin 2003-02-09 16:02:00 UTC
I agree that it is most likely a gcc issue.  I *think* gcc may have failed the sanity test.  No gcc ended up on my $PATH.  I know that cpp consistently fails the sanity check.  I think this is related to bug 12699.

I have the failed system intact on a partition on my hard drive if that helps. I can chroot into it from a working Gentoo and I am willing to do whatever you suggest to try to track this down.

I thank you all for your rapid response.  I need go out this after noon (PST), but I will be back tonight ready to experiment.

Comment 8 SpanKY gentoo-dev 2003-02-09 16:09:44 UTC
ill tackle each of your issues ;) 
Comment 9 SpanKY gentoo-dev 2003-02-09 16:13:48 UTC
the access denied is an old portage bug ... update portage to fix it ... 
 
point 2: 
depends on the ebuild, not all packages respect CFLAGS so sometimes we have to 
hack them ... please open a new bug about any offending packages ... 
 
point 3: 
none of the tbz2's are corrupted, they're portage binary packages, not just bzipped 
tarballs ... that means we append some data at the end (trailing garbage) 
 
point 6: 
you may have had some directories still mounted (/mnt/gentoo/proc or 
/mnt/gentoo/dev), or you had some shells sitting in /mnt/gentoo ... 
oh, and although umount provides a '-f' option, it doesnt work, like ever :) 
 
point 7: 
you cannot currently do multiple portage commands at the sametime and not have 
them break ... portage does not check to make sure it's not already emerging the 
package before going ahead and attempting it 
 
i think we covered everything ;) 

*** This bug has been marked as a duplicate of 13013 ***
Comment 10 James Bowlin 2003-02-09 23:23:09 UTC
Thanks for all of your help.   
 
I emerged gcc (after portage) because the failed sanity checks  
were the first problem I encountered, I am still having problems 
with sanity checks while emerging gcc. I think my problem is related 
to the still open 12699.    
 
I will be glad to do whatever you suggest.  
 
I did a rm -Rf /mnt/gentoo/*  and then: 
 
  523  tar -xjpf /mnt/cdrom/gentoo/stage3-pentium3-1.4_rc2.tbz2 
  524  mount -o bind /proc /mnt/gentoo/proc 
  526  cp /etc/resolv.conf /mnt/gentoo/etc/ 
  527  chroot /mnt/gentoo /bin/bash 
    1  env-update 
   15  source /etc/profile 
   16  emerge sync 
   17  emerge portage 
   45  emerge -u sys-devel/gcc-config 
   47  emerge -u sys-devel/gcc 
 
I was not watching carefully for error messages on the console window. 
I think I still have sanity check problems because after emerging gcc: 
 
   1) /usr/bin/cpp was converted from a binary to a script. 
   2) /usr/bin/gcc disappeared 
 
before:  
    94408 Dec  4 19:45 cpp* 
    93448 Dec  4 19:45 gcc* 
after: 
        303 Feb  9 18:47 cpp 
        [no gcc] 
 
# emerge info 
Portage 2.0.46-r12 (default-x86-1.4, gcc-3.2.1, glibc-2.3.1-r2) 
================================================================= 
System uname: 2.4.19-gentoo-r10 i686 Intel(R) Pentium(R) 4 Mobile CPU 1.60GHz 
GENTOO_MIRRORS="http://gentoo.oregonstate.edu 
http://csociety-ftp.ecn.purdue.edu/pub/gentoo 
http://www.ibiblio.org/pub/Linux/distributions/gentoo" 
CONFIG_PROTECT="/etc /var/qmail/control /usr/kde/2/share/config /usr/kde/3.1/share/config 
/usr/kde/3/share/config /usr/share/config" 
CONFIG_PROTECT_MASK="/etc/gconf /etc/env.d" 
PORTDIR="/usr/portage" 
DISTDIR="/usr/portage/distfiles" 
PKGDIR="/usr/portage/packages" 
PORTAGE_TMPDIR="/var/tmp" 
PORTDIR_OVERLAY="" 
USE="x86 oss apm arts avi crypt cups gif gpm gtk imlib java jpeg kde libg++ libwww mikmod 
mmx motif mpeg ncurses nls oggvorbis opengl pdflib png qt qtmt quicktime sdl spell svga 
truetype X xml2 xmms xv zlib gdbm berkdb slang readline tcpd pam ssl python -3dnow atlas 
cdr dga dvd -encode gd ggz -gnome jikes mbox mozilla oav pcmcia perl plotutils pnp ruby sse 
tcltk tiff usb xml" 
COMPILER="gcc3" 
CHOST="i686-pc-linux-gnu" 
CFLAGS="-O3 -march=pentium4 -fomit-frame-pointer -falign-functions=4 -pipe" 
CXXFLAGS="-O3 -march=pentium4 -fomit-frame-pointer -falign-functions=4 -pipe" 
ACCEPT_KEYWORDS="x86" 
MAKEOPTS="-j2" 
AUTOCLEAN="yes" 
SYNC="rsync://rsync.gentoo.org/gentoo-portage" 
FEATURES="sandbox ccache" 
 
 
 
 
Comment 11 Seemant Kulleen (RETIRED) gentoo-dev 2003-02-09 23:42:21 UTC
yeah, gcc moves to a different location -- do: which gcc
at a shell prompt to see its new location
Comment 12 James Bowlin 2003-02-10 03:39:16 UTC
Things are still broken.  After the:

# emerge sync
# emerge portage
# emerge sys-devel/gcc

I continued as if nothing were wrong:

# emerge -u world

I got the following error message when emerge aborted:

!!! ERROR: x11-base/xfree-4.2.1-r2 failed.
!!! Function src_install, line 211, Exitcode 2
!!! (no error message)

In summary, when I:

# emerge sync
# emerge portage
# emerge -u world

The following is happening:

1) I get sanity check failure(s) while emerging gcc.
2) I then get ACCESS DENIED errors while emerging TrueType fonts in xfree
3) The xfree emerge finally aborts with the above message

The problems are consistent and reproducible.  I've done it 3 times
now, once without emerge portage and twice with it.

Comment 13 Seemant Kulleen (RETIRED) gentoo-dev 2003-02-10 03:46:30 UTC
can you paste the access errors, as well as emerge --version?
Comment 14 James Bowlin 2003-02-10 04:56:38 UTC
# emerge --version 
Portage 2.0.46-r12 (default-x86-1.4, gcc-3.2.1, glibc-2.3.1-r2) 
 
One of the previous access errors is in the first report above. 
My emerges take many hours and I can't sit in front of the 
screen anymore waiting to capture error messages. 
 
Is there a standard way to collect the output of emerge in a file? 
 
If so, would you like me to emerge again and capture the output 
so I can find the error messages? 
 
Also, is there a faster way to get the system back into the 
state just before emerging? Currently I have been doing: 
 
# rm -Rf /mnt/gentoo/* 
# cd /mnt/gentoo 
# tar xjpf /mnt/cdrom/gentoo/stage3-pentium3-1.4_rc2.tbz2 
 
 
Comment 15 Seemant Kulleen (RETIRED) gentoo-dev 2003-02-10 06:34:43 UTC
to log:

emerge something &>~/something.emerge.log

then just attach those logs (placed in your home dir in this case) to this bug report.
Comment 16 Seemant Kulleen (RETIRED) gentoo-dev 2003-02-12 09:19:45 UTC
still with me jim?
Comment 17 Nicholas Jones (RETIRED) gentoo-dev 2003-02-12 10:07:02 UTC
Look in make.conf for: PORT_LOGDIR=/var/log/emerge

mtimedb/Access-Violation issues are caused by older portage versions.

pentium4 is _known_ to cause problems. Don't use it.

The script over gcc is intentional. source /etc/profile to fix missing gcc's
and use gcc-config to set the one you want.

If there's anything that isn't covered or recurs after you remove pentium4,
then feel free to reopen.
Comment 18 James Bowlin 2003-02-16 01:53:35 UTC
Thank you all for your help.  I am now pretty sure that I was suffering from a 
temperature related hardware problem.  I was able to install Gentoo on 
ext2 but then I had 4 (or more?) failures trying to install on reiserfs.  Their 
website says that rieser can cause significantly more heating (I guess because 
it is so much faster).  Anyway, I get i8k working (on my i8200) and manually  
set both fans to high.  The installation on the reiserfs then worked without 
a hitch.  Thanks again. 
Comment 19 John Davis (zhen) (RETIRED) gentoo-dev 2003-04-04 01:22:42 UTC
db fix
Comment 20 John Davis (zhen) (RETIRED) gentoo-dev 2003-04-04 01:27:34 UTC
db fix