[lxc-users] Question about existing defaults
CDR
venefax at gmail.com
Fri Apr 18 04:33:15 UTC 2014
This may help you understand the issue
yum whatprovides "*/caps.c"
Loaded plugins: langpacks, refresh-packagekit
kcbench-data-3.9-0.1-14.fc20.noarch : Kernel sources from 3.9 to be
used by kcbench
Repo : fedora
Matched from:
Filename : /usr/share/kcbench-data/linux-3.9/drivers/net/wireless/ath/ath5k/caps.c
Filename : /usr/share/kcbench-data/linux-3.9/fs/ceph/caps.c
lxc-debuginfo-1.0.3-1.fc20.x86_64 : Debug information for package lxc
Repo : @/lxc-debuginfo-1.0.3-1.fc20.x86_64
Matched from:
Filename : /usr/src/debug/lxc-1.0.3/src/lxc/caps.c
On Fri, Apr 18, 2014 at 12:30 AM, CDR <venefax at gmail.com> wrote:
> I can give you direct access to the box where I am developing this.
> Please write me to my email directly.
> There is nothing strange about my Fedora20 box. It is fully yum-updated.
>
> ldd /usr/bin/lxc-start
> linux-vdso.so.1 => (0x00007fff0d7d1000)
> liblxc.so.1 => /lib/liblxc.so.1 (0x0000003d98c00000)
> libcap.so.2 => /lib64/libcap.so.2 (0x00000039d1600000)
> libutil.so.1 => /lib64/libutil.so.1 (0x00000039e6600000)
> libpthread.so.0 => /lib64/libpthread.so.0 (0x00000039bba00000)
> libc.so.6 => /lib64/libc.so.6 (0x00000039bb200000)
> /lib64/ld-linux-x86-64.so.2 (0x00000039bae00000)
> libattr.so.1 => /lib64/libattr.so.1 (0x00000039c8600000)
>
> On Thu, Apr 17, 2014 at 10:22 PM, Michael H. Warfield <mhw at wittsend.com> wrote:
>> On Thu, 2014-04-17 at 21:25 -0400, CDR wrote:
>>> Dear friends
>>> I installed the RPMs again, and rebooted. Previous, I issued an
>>> "ldconfig", just in case. It still fails.
>>> Maybe one of the developers of lxc may take a look at the trace below
>>> and spot the issue. The advantages of the current version over the one
>>> issued in Fedora are too great to ignore, like "lxc-autostart", etc.
>>
>> This is not occurring on my F20 host and I'm rotating development and
>> release versions almost on a daily basis. This isn't making any sense.
>> That function should exist. It does exist on a clean install. It's
>> defined in caps.c.
>>
>> What do you get from this:
>>
>> ldd /usr/bin/lxc-start
>>
>> You seem to still be hooking up with the wrong library for some reason.
>>
>> Regards,
>> Mike
>>
>>> strace lxc-start -d -n dialer-1
>>> execve("/usr/bin/lxc-start", ["lxc-start", "-d", "-n", "dialer-1"],
>>> [/* 31 vars */]) = 0
>>> brk(0) = 0x604000
>>> mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,
>>> 0) = 0x7fb2a9a35000
>>> access("/etc/ld.so.preload", R_OK) = -1 ENOENT (No such file or directory)
>>> open("/etc/ld.so.cache", O_RDONLY|O_CLOEXEC) = 3
>>> fstat(3, {st_mode=S_IFREG|0644, st_size=103464, ...}) = 0
>>> mmap(NULL, 103464, PROT_READ, MAP_PRIVATE, 3, 0) = 0x7fb2a9a1b000
>>> close(3) = 0
>>> open("/lib/liblxc.so.1", O_RDONLY|O_CLOEXEC) = 3
>>> read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\200\313\300\230=\0\0\0"...,
>>> 832) = 832
>>> fstat(3, {st_mode=S_IFREG|0755, st_size=1497976, ...}) = 0
>>> mmap(0x3d98c00000, 2488264, PROT_READ|PROT_EXEC,
>>> MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x3d98c00000
>>> mprotect(0x3d98c5e000, 2093056, PROT_NONE) = 0
>>> mmap(0x3d98e5d000, 12288, PROT_READ|PROT_WRITE,
>>> MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x5d000) = 0x3d98e5d000
>>> close(3) = 0
>>> open("/lib64/libcap.so.2", O_RDONLY|O_CLOEXEC) = 3
>>> read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0
>>> \26`\3219\0\0\0"..., 832) = 832
>>> fstat(3, {st_mode=S_IFREG|0755, st_size=21424, ...}) = 0
>>> mmap(0x39d1600000, 2114112, PROT_READ|PROT_EXEC,
>>> MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x39d1600000
>>> mprotect(0x39d1604000, 2093056, PROT_NONE) = 0
>>> mmap(0x39d1803000, 8192, PROT_READ|PROT_WRITE,
>>> MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x3000) = 0x39d1803000
>>> close(3) = 0
>>> open("/lib64/libutil.so.1", O_RDONLY|O_CLOEXEC) = 3
>>> read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\20\17`\3469\0\0\0"...,
>>> 832) = 832
>>> fstat(3, {st_mode=S_IFREG|0755, st_size=17536, ...}) = 0
>>> mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,
>>> 0) = 0x7fb2a9a1a000
>>> mmap(0x39e6600000, 2105616, PROT_READ|PROT_EXEC,
>>> MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x39e6600000
>>> mprotect(0x39e6602000, 2093056, PROT_NONE) = 0
>>> mmap(0x39e6801000, 8192, PROT_READ|PROT_WRITE,
>>> MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x1000) = 0x39e6801000
>>> close(3) = 0
>>> open("/lib64/libpthread.so.0", O_RDONLY|O_CLOEXEC) = 3
>>> read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\340m\240\2739\0\0\0"...,
>>> 832) = 832
>>> fstat(3, {st_mode=S_IFREG|0755, st_size=150800, ...}) = 0
>>> mmap(0x39bba00000, 2213104, PROT_READ|PROT_EXEC,
>>> MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x39bba00000
>>> mprotect(0x39bba18000, 2093056, PROT_NONE) = 0
>>> mmap(0x39bbc17000, 8192, PROT_READ|PROT_WRITE,
>>> MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x17000) = 0x39bbc17000
>>> mmap(0x39bbc19000, 13552, PROT_READ|PROT_WRITE,
>>> MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x39bbc19000
>>> close(3) = 0
>>> open("/lib64/libc.so.6", O_RDONLY|O_CLOEXEC) = 3
>>> read(3, "\177ELF\2\1\1\3\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0p\36\"\2739\0\0\0"...,
>>> 832) = 832
>>> fstat(3, {st_mode=S_IFREG|0755, st_size=2100672, ...}) = 0
>>> mmap(0x39bb200000, 3924576, PROT_READ|PROT_EXEC,
>>> MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x39bb200000
>>> mprotect(0x39bb3b4000, 2097152, PROT_NONE) = 0
>>> mmap(0x39bb5b4000, 24576, PROT_READ|PROT_WRITE,
>>> MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x1b4000) = 0x39bb5b4000
>>> mmap(0x39bb5ba000, 16992, PROT_READ|PROT_WRITE,
>>> MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x39bb5ba000
>>> close(3) = 0
>>> open("/lib64/libattr.so.1", O_RDONLY|O_CLOEXEC) = 3
>>> read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\320\23`\3109\0\0\0"...,
>>> 832) = 832
>>> fstat(3, {st_mode=S_IFREG|0755, st_size=22160, ...}) = 0
>>> mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,
>>> 0) = 0x7fb2a9a19000
>>> mmap(0x39c8600000, 2113904, PROT_READ|PROT_EXEC,
>>> MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x39c8600000
>>> mprotect(0x39c8604000, 2093056, PROT_NONE) = 0
>>> mmap(0x39c8803000, 8192, PROT_READ|PROT_WRITE,
>>> MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x3000) = 0x39c8803000
>>> close(3) = 0
>>> mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,
>>> 0) = 0x7fb2a9a18000
>>> mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,
>>> 0) = 0x7fb2a9a16000
>>> arch_prctl(ARCH_SET_FS, 0x7fb2a9a167c0) = 0
>>> mprotect(0x39bb5b4000, 16384, PROT_READ) = 0
>>> mprotect(0x39c8803000, 4096, PROT_READ) = 0
>>> mprotect(0x39bbc17000, 4096, PROT_READ) = 0
>>> mprotect(0x39e6801000, 4096, PROT_READ) = 0
>>> mprotect(0x39d1803000, 4096, PROT_READ) = 0
>>> mprotect(0x3d98e5d000, 4096, PROT_READ) = 0
>>> mprotect(0x602000, 4096, PROT_READ) = 0
>>> mprotect(0x39bb01f000, 4096, PROT_READ) = 0
>>> munmap(0x7fb2a9a1b000, 103464) = 0
>>> set_tid_address(0x7fb2a9a16a90) = 3388
>>> set_robust_list(0x7fb2a9a16aa0, 24) = 0
>>> rt_sigaction(SIGRTMIN, {0x39bba068c0, [], SA_RESTORER|SA_SIGINFO,
>>> 0x39bba0f750}, NULL, 8) = 0
>>> rt_sigaction(SIGRT_1, {0x39bba06950, [],
>>> SA_RESTORER|SA_RESTART|SA_SIGINFO, 0x39bba0f750}, NULL, 8) = 0
>>> rt_sigprocmask(SIG_UNBLOCK, [RTMIN RT_1], NULL, 8) = 0
>>> getrlimit(RLIMIT_STACK, {rlim_cur=8192*1024, rlim_max=RLIM64_INFINITY}) = 0
>>> writev(2, [{"lxc-start", 9}, {": ", 2}, {"symbol lookup error", 19},
>>> {": ", 2}, {"lxc-start", 9}, {": ", 2}, {"undefined symbol:
>>> lxc_caps_init", 31}, {"", 0}, {"", 0}, {"\n", 1}], 10lxc-start: symbol
>>> lookup error: lxc-start: undefined symbol: lxc_caps_init
>>> ) = 75
>>> exit_group(127) = ?
>>> +++ exited with 127 +++
>>>
>>> On Thu, Apr 17, 2014 at 2:27 PM, Michael H. Warfield <mhw at wittsend.com> wrote:
>>> > On Thu, 2014-04-17 at 07:29 -0400, CDR wrote:
>>> >> I followed the instructions and it works, indeed.
>>> >> I erased first all lxc rpms and the installed the new ones
>>> >
>>> >> rpm -qa |grep lxc
>>> >> lxc-libs-1.0.3-1.fc20.x86_64
>>> >> lxc-debuginfo-1.0.3-1.fc20.x86_64
>>> >> lxc-1.0.3-1.fc20.x86_64
>>> >> libvirt-daemon-driver-lxc-1.1.3.4-4.fc20.x86_64
>>> >
>>> >> However, I hit a problem when starting a container.
>>> >> lxc-start -d -n dialer-1
>>> >> lxc-start: symbol lookup error: lxc-start: undefined symbol: lxc_caps_init
>>> >
>>> >> This means that the code is left to pull this function from package
>>> >> that is not there?
>>> >
>>> > That usually means you've got an old library dangling around or you have
>>> > processes with a deleted library locked in memory. That most often
>>> > happens when you've done a manual "./configure ; make ; make install"
>>> > and pieces are now installed where you don't want them. In that case,
>>> > ldconfig may have the lxc-libs located in a different path location than
>>> > what you've installed. It could also occur if you have running
>>> > containers with an incompatible library.
>>> >
>>> > First thing I would do is inspect /usr/local and look for any lxc
>>> > components and get rid of them.
>>> >
>>> > Second thing I would do is reboot to make sure you have no stale
>>> > processes holding an old library in common memory.
>>> >
>>> > Then try it again.
>>> >
>>> > This is why I strictly go with the yum/rpm route when working with
>>> > managed packages like LXC.
>>> >
>>> >> Any help would be appreciated, or a patch that would not use this
>>> >> function. My containers are unbound, only for internal use, so I want
>>> >> them to use as much resources as needed.
>>> >
>>> > What rpm is showing is that you have a coherent set of packages that
>>> > should not be encountering this problem. The error indicates the
>>> > presence of an unmanaged library in memory on on disk. You might track
>>> > it down using "ldd /usr/bin/lxc-start" and see what that gives you.
>>> >
>>> >> Philip
>>> >
>>> > Regards,
>>> > Mike
>>> >
>>> >> On Wed, Apr 16, 2014 at 11:41 PM, Michael H. Warfield <mhw at wittsend.com> wrote:
>>> >> > On Wed, 2014-04-16 at 22:04 -0400, CDR wrote:
>>> >> >> I am using Fedora 20 and try to compile the very latest sources,
>>> >> >> because they bring new utilities that are very useful.
>>> >> >> You would assume, I think, that by using ./configure
>>> >> >> --with-distro=fedora, the right defaults would be pulled over. It is
>>> >> >> not the case. I need to know how to apply the existing defaults if the
>>> >> >> software is already running in the box, so the libraries and utilities
>>> >> >> do not get written to a different places. Particularly annoying is the
>>> >> >> location of the config files.
>>> >> >
>>> >> >> In summary, don't we agree that by using the distro name, the software
>>> >> >> should either extract or know the right defaults? Unless somebody has
>>> >> >> a secret way to extract that information from the RPM offered by the
>>> >> >> distribution.
>>> >> >
>>> >> > Interesting. I guess this is one I might field, since I'm responsible
>>> >> > for the Fedora template (which does NOT imply the Fedora build). I
>>> >> > don't know who the Fedora package manager for the LXC package is and I
>>> >> > haven't heard from him or her.
>>> >> >
>>> >> > I, for one, would NOT expect "./configure --with-distro=fedora" to pull
>>> >> > the correct results. That's from experience, although I've never used
>>> >> > it myself. It's just something none of us have thought to maintain and
>>> >> > I have never looked into it. I didn't even know the option was there in
>>> >> > configure.
>>> >> >
>>> >> > The Fedora Project itself does not even use the default lxc.spec rpm
>>> >> > spec file from the project but chooses their own .spec file and
>>> >> > packaging. Largely, the defaults are similar...
>>> >> >
>>> >> > I had been using an explicit build based on our .spec file to create tar
>>> >> > files and build rpm files using "rpmbuild -ta" but another contributor,
>>> >> > Dwight (Oracle), pointed out to me that I could just do a "make rpm" in
>>> >> > a source directory and it would to the correct things to create a Fedora
>>> >> > rpm. I've been very successful doing that and now that's all I do. It
>>> >> > does the right things. It uses slightly different packaging from the
>>> >> > "official" Fedora packages but one that's compatible with the Fedora
>>> >> > packaging.
>>> >> >
>>> >> > If you are manually running "./configure" and "make" on our sources on
>>> >> > Fedora, my first advice would be DON'T. You're doing things the hard
>>> >> > way. I don't even do that with any stock rpm packages I'm working on.
>>> >> > Work through the yum/rpm system. Run "make rpm" and install the rpms
>>> >> > using "yum localinstall". That will prevent you from ending up with
>>> >> > conflicting binaries in differing locations.
>>> >> >
>>> >> > In this case, it's actually easier to do the right thing than it is to
>>> >> > do the manual thing the hard way. You're trying too hard.
>>> >> >
>>> >> >> Philip
>>> >> >
>>> >> > Regards,
>>> >> > Mike
>>> >> > --
>>> >> > Michael H. Warfield (AI4NB) | (770) 978-7061 | mhw at WittsEnd.com
>>> >> > /\/\|=mhw=|\/\/ | (678) 463-0932 | http://www.wittsend.com/mhw/
>>> >> > NIC whois: MHW9 | An optimist believes we live in the best of all
>>> >> > PGP Key: 0x674627FF | possible worlds. A pessimist is sure of it!
>>> >> >
>>> >> >
>>> >> > _______________________________________________
>>> >> > lxc-users mailing list
>>> >> > lxc-users at lists.linuxcontainers.org
>>> >> > http://lists.linuxcontainers.org/listinfo/lxc-users
>>> >> _______________________________________________
>>> >> lxc-users mailing list
>>> >> lxc-users at lists.linuxcontainers.org
>>> >> http://lists.linuxcontainers.org/listinfo/lxc-users
>>> >
>>> > --
>>> > Michael H. Warfield (AI4NB) | (770) 978-7061 | mhw at WittsEnd.com
>>> > /\/\|=mhw=|\/\/ | (678) 463-0932 | http://www.wittsend.com/mhw/
>>> > NIC whois: MHW9 | An optimist believes we live in the best of all
>>> > PGP Key: 0x674627FF | possible worlds. A pessimist is sure of it!
>>> >
>>> >
>>> > _______________________________________________
>>> > lxc-users mailing list
>>> > lxc-users at lists.linuxcontainers.org
>>> > http://lists.linuxcontainers.org/listinfo/lxc-users
>>> _______________________________________________
>>> lxc-users mailing list
>>> lxc-users at lists.linuxcontainers.org
>>> http://lists.linuxcontainers.org/listinfo/lxc-users
>>
>> --
>> Michael H. Warfield (AI4NB) | (770) 978-7061 | mhw at WittsEnd.com
>> /\/\|=mhw=|\/\/ | (678) 463-0932 | http://www.wittsend.com/mhw/
>> NIC whois: MHW9 | An optimist believes we live in the best of all
>> PGP Key: 0x674627FF | possible worlds. A pessimist is sure of it!
>>
>>
>> _______________________________________________
>> lxc-users mailing list
>> lxc-users at lists.linuxcontainers.org
>> http://lists.linuxcontainers.org/listinfo/lxc-users
More information about the lxc-users
mailing list