Linux Development – Page 16

Poor Yum Repo: OracleLinux6

I’m now working on an Oracle Linux 6U1 x64 box, planning to inst new 11R2 image.

Problem hit here.

# yum install unixODBC-devel.i686
...
--&gt; Running transaction check
---&gt; Package unixODBC.i686 0:2.2.14-11.el6 will be installed
--&gt; Processing Dependency: libc.so.6(GLIBC_2.7) for package: unixODBC-2.2.14-11.el6.i686
...
---&gt; Package glibc.i686 0:2.12-1.25.el6 will be installed
...--&gt; Processing Dependency: glibc-common = 2.12-1.25.el6 for package: glibc-2.12-1.25.el6.i686
---&gt; Package ncurses-libs.i686 0:5.7-3.20090208.el6 will be installed
---&gt; Package nss-softokn-freebl.i686 0:3.12.9-3.el6 will be installed
--&gt; Finished Dependency Resolution
Error: Package: glibc-2.12-1.25.el6.i686 (ol6_u1_base)
           Requires: glibc-common = 2.12-1.25.el6
           Installed: glibc-common-2.12-1.25.el6_1.3.x86_64 (@ol6_latest)
               glibc-common = 2.12-1.25.el6_1.3
           Available: glibc-common-2.12-1.25.el6.x86_64 (ol6_u1_base)
               glibc-common = 2.12-1.25.el6
 You could try using --skip-broken to work around the problem
 You could try running: rpm -Va --nofiles --nodigest

# yum install unixODBC-devel.i686

...

--> Running transaction check

---> Package unixODBC.i686 0:2.2.14-11.el6 will be installed

--> Processing Dependency: libc.so.6(GLIBC_2.7) for package: unixODBC-2.2.14-11.el6.i686

...

---> Package glibc.i686 0:2.12-1.25.el6 will be installed

...--> Processing Dependency: glibc-common = 2.12-1.25.el6 for package: glibc-2.12-1.25.el6.i686

---> Package ncurses-libs.i686 0:5.7-3.20090208.el6 will be installed

---> Package nss-softokn-freebl.i686 0:3.12.9-3.el6 will be installed

--> Finished Dependency Resolution

Error: Package: glibc-2.12-1.25.el6.i686 (ol6_u1_base)

Requires: glibc-common = 2.12-1.25.el6

Installed: glibc-common-2.12-1.25.el6_1.3.x86_64 (@ol6_latest)

glibc-common = 2.12-1.25.el6_1.3

Available: glibc-common-2.12-1.25.el6.x86_64 (ol6_u1_base)

glibc-common = 2.12-1.25.el6

You could try using --skip-broken to work around the problem

You could try running: rpm -Va --nofiles --nodigest

we can see that glibc.i686 <tip: just add arch. tag to install related pkg.> cannot be resolved for glibc-common label check failure.

it’s funny but not worse, ORCL even doesn’t provide glibc-common.i686 package…
http://public-yum.oracle.com/repo/OracleLinux/OL6/1/base/x86_64/

It should be a bug here….

WorkAround:

1. pull glibc package from weburl glibc-2.12-1.25.el6.i686.rpm.
2. # yum localinstall glibc-2.12-1.25.el6.i686.rpm
3. # yum install unixODBC-devel unixODBC-devel.i686
4. continue…

Notes….
Poor Yum repo…especially on x64 platform day’s, mixing with i686 libs.

Emerald package has retired

It’s so long time to issue yum-update on my F15 boxes.
found the emerald packages already removed from repo…

[Following links tell everything:]
https://bugzilla.redhat.com/show_bug.cgi?id=739617
http://fedoraforum.org/forum/showthread.php?p=1516387
http://pkgs.fedoraproject.org/gitweb/?p=emerald.git;a=shortlog;h=refs/heads/master

[Workaround to continue running Emerald:]
just go back to 0.9.4~
1. Download essential packages:
[x86_64]

wget http://mirrors.reflected.net/fedora/linux/releases/15/Everything/x86_64/os/Packages/{compiz-0.9.4-1.fc15.x86_64.rpm,compiz-gtk-0.9.4-1.fc15.x86_64.rpm,ccsm-0.9.4-1.fc15.noarch.rpm,compiz-plugins-main-0.9.4-1.fc15.x86_64.rpm,compizconfig-python-0.9.4-1.fc15.x86_64.rpm,libcompizconfig-0.9.4-1.fc15.x86_64.rpm,emerald-0.8.5-0.4.git90c9604441.fc15.x86_64.rpm,emerald-themes-0.5.2-6.fc15.noarch.rpm }

wget http://mirrors.reflected.net/fedora/linux/releases/15/Everything/x86_64/os/Packages/{compiz-0.9.4-1.fc15.x86_64.rpm,compiz-gtk-0.9.4-1.fc15.x86_64.rpm,ccsm-0.9.4-1.fc15.noarch.rpm,compiz-plugins-main-0.9.4-1.fc15.x86_64.rpm,compizconfig-python-0.9.4-1.fc15.x86_64.rpm,libcompizconfig-0.9.4-1.fc15.x86_64.rpm,emerald-0.8.5-0.4.git90c9604441.fc15.x86_64.rpm,emerald-themes-0.5.2-6.fc15.noarch.rpm }

[i386]

wget http://mirrors.reflected.net/fedora/linux/releases/15/Everything/i386/os/Packages/{compiz-0.9.4-1.fc15.i686.rpm,compiz-gtk-0.9.4-1.fc15.i686.rpm,ccsm-0.9.4-1.fc15.noarch.rpm,compiz-plugins-main-0.9.4-1.fc15.i686.rpm,compizconfig-python-0.9.4-1.fc15.i686.rpm,libcompizconfig-0.9.4-1.fc15.i686.rpm,emerald-0.8.5-0.4.git90c9604441.fc15.i686.rpm,emerald-themes-0.5.2-6.fc15.noarch.rpm}

wget http://mirrors.reflected.net/fedora/linux/releases/15/Everything/i386/os/Packages/{compiz-0.9.4-1.fc15.i686.rpm,compiz-gtk-0.9.4-1.fc15.i686.rpm,ccsm-0.9.4-1.fc15.noarch.rpm,compiz-plugins-main-0.9.4-1.fc15.i686.rpm,compizconfig-python-0.9.4-1.fc15.i686.rpm,libcompizconfig-0.9.4-1.fc15.i686.rpm,emerald-0.8.5-0.4.git90c9604441.fc15.i686.rpm,emerald-themes-0.5.2-6.fc15.noarch.rpm}

2. Remove current 0.9.5.0 compiz related packages.

rpm -qa | grep 0.9.5.0 | xargs rpm -e --nodeps

1	rpm -qa \| grep 0.9.5.0 \| xargs rpm -e --nodeps

3. install the RPMs downloaded above.

4. edit the /etc/yum.conf to block compiz/emerald related updates.

 echo "exclude=ccsm* compiz* compizconfig-python* compiz-gtk* compiz-plugins-main* libcompizconfig*" &gt;&gt; /etc/yum.conf

1	echo "exclude=ccsm* compiz* compizconfig-python* compiz-gtk* compiz-plugins-main* libcompizconfig*" >> /etc/yum.conf

Well now I’ve blocked updates on compiz and libreoffice… is there any more?

VMSetError: VirtualBox can’t operate in VMX root mode.

VirtualBox failed to start VM guest on OEL5 node,

00:00:00.186 Package type: LINUX_64BITS_RHEL_5
...
...
00:00:00.190 VMSetError: VirtualBox can't operate in VMX root mode. Please disable the KVM kernel extension, recompile your kernel and reboot
00:00:00.190 ERROR [COM]: aRC=NS_ERROR_FAILURE (0x80004005) aIID={515e8e8d-f932-4d8e-9f32-79a52aead882} aComponent={Console} 
             aText={VirtualBox can't operate in VMX root mode. Please disable the KVM kernel extension, recompile your kernel and 
                   reboot (VERR_VMX_IN_VMX_ROOT_MODE)}, 
             preserve=false
00:00:00.192 Using known keycode mapping for keycode to scan code conversion
00:00:00.193 Power up failed (vrc=VERR_VMX_IN_VMX_ROOT_MODE, rc=NS_ERROR_FAILURE (0X80004005))

00:00:00.186 Package type: LINUX_64BITS_RHEL_5

...

00:00:00.190 VMSetError: VirtualBox can't operate in VMX root mode. Please disable the KVM kernel extension, recompile your kernel and reboot

00:00:00.190 ERROR [COM]: aRC=NS_ERROR_FAILURE (0x80004005) aIID={515e8e8d-f932-4d8e-9f32-79a52aead882} aComponent={Console}

aText={VirtualBox can't operate in VMX root mode. Please disable the KVM kernel extension, recompile your kernel and

reboot (VERR_VMX_IN_VMX_ROOT_MODE)},

preserve=false

00:00:00.192 Using known keycode mapping for keycode to scan code conversion

00:00:00.193 Power up failed (vrc=VERR_VMX_IN_VMX_ROOT_MODE, rc=NS_ERROR_FAILURE (0X80004005))

To Workaround:

# modprobe -r kvm-intel

1	# modprobe -r kvm-intel

chmsee doesnt work on Fedora 15 official

just found that the chmsee tool doesnt work well today. it should worked at the release version and failed when yum up2date.
check and filed a bug on this issue.

[suse@suse ~]$LANG=
[suse@suse ~]$chmsee

** (chmsee:17763): WARNING **: GECKO_UTILS &gt;&gt;&gt; Couldn't find a compatible GRE!

** ERROR **: Initialize html render engine failed!
aborting...
Aborted

[suse@suse ~]$LANG=

[suse@suse ~]$chmsee

** (chmsee:17763): WARNING **: GECKO_UTILS >>> Couldn't find a compatible GRE!

** ERROR **: Initialize html render engine failed!

aborting...

Aborted

W/R
export GRE_HOME env before run the tool,

<pre class="brush:shell">$export GRE_HOME=/usr/lib/xulrunner-2/
$chmsee

1 2	<pre class="brush:shell">$export GRE_HOME=/usr/lib/xulrunner-2/ $chmsee

compile rose-0.9.5a-14791 on Fedora 15

rose is a nice tool, suggested when reading the book <The CERT C Secure Coding Standard>

just install the tool on my Fedora 15 node,

1. Fetch the latest source code from official website: 0.9.5a-multiplatform-14791
2. Prepare the packages needed,(I use official Java Bins configured, which can be detected by configure script)

yum install  boost-devel  gcc-gfortran

1	yum install boost-devel gcc-gfortran

3. modify the configure script incase a boost version detect failure.

8622c8622
&lt; rose_boost_version=`grep "#define BOOST_VERSION " ${ac_boost_path}/include/boost/version.hpp | cut -d" " -f 3`
---
&gt; rose_boost_version=`grep "#define BOOST_VERSION " /usr/include/boost/version.hpp | cut -d" " -f 3`

8622c8622

< rose_boost_version=`grep "#define BOOST_VERSION " ${ac_boost_path}/include/boost/version.hpp | cut -d" " -f 3`

---

> rose_boost_version=`grep "#define BOOST_VERSION " /usr/include/boost/version.hpp | cut -d" " -f 3`

**About the path , I use following line to determine, it may also works on RH related platform.

      $grep BOOST_VERSION `rpm -ql boost-devel| grep version`

1	$grep BOOST_VERSION `rpm -ql boost-devel\| grep version`

4. configure with setting, for currently fedora.15 using gcc-4.6.0.

--disable-gcc-version-check
         Disable GCC version 4.0.x - 4.4.x verification check

1 2	--disable-gcc-version-check Disable GCC version 4.0.x - 4.4.x verification check

If you do only need some(not all) languages supported, using

--enable-languages=LIST Build specific languages:
                 all,none,binary-analysis,c,c++,cuda,fortran,java,opencl,php
                 (default=all)

--enable-languages=LIST Build specific languages:

all,none,binary-analysis,c,c++,cuda,fortran,java,opencl,php

(default=all)

5. make and continue.
Before make do modify the Makefile, or the script shall not fetch the related EDG binaries for version4.6.

320c320
&lt; GENERATED_COMPILER_NAME_AND_VERSION_SUFFIX = GNU-4.6
---
&gt; GENERATED_COMPILER_NAME_AND_VERSION_SUFFIX = GNU-4.4

320c320

< GENERATED_COMPILER_NAME_AND_VERSION_SUFFIX = GNU-4.6

---

> GENERATED_COMPILER_NAME_AND_VERSION_SUFFIX = GNU-4.4