April 2008 - Devel - libvirt List Archives

[Libvir] [PATCH 1/2] lxc: start container
by Dave Leskovec 10 Apr '08

10 Apr '08

This is a repost of the start container support. Changes from the last version: * Report an error when allocation for init string fails in lxcExecContainerInit() * Change to find by name in lxcStartDomain() * Move tty forwarding process outside of the container. This allows consolidating the forwarding into a single process at a later time. This also means the the container init process as specified by the user now runs as the container root process with pid = 1. The tty setup will require some (hopefully minor) modifications when pts namespaces are enabled. * Add header comments to a number of the functions. This is an updated rough outline of the functions involved in starting a container and the namespace and process under which they run: lxcVmStart() - runs under libvirtd process lxcSetupTtyTunnel() - opens and configures parent tty lxcSetupContainerTty() - opens container tty fork child process calls lxcTtyForward() see below parent continues lxcStartContainer - see below return lxcStartContainer() - runs in parent namespace, libvirtd process Allocate stack for container clone() - child process will start in lxcChild() see below return lxcChild() - runs within container, child process from clone() mount user filesystems mount container /proc lxcExecWithTty() - see below, will not return lxcExecWithTty() - runs within container, root process lxcSetContainerStdio - sets container tty as primary console lxcExecContainerInit - see below, should not return exit() lxcExecContainerInit() - runs within container, root process exec containers init if exec fails, exit() Thanks! -- Best Regards, Dave Leskovec IBM Linux Technology Center Open Virtualization

2 2

[Libvir] [PATCH REPOST] DHCP host mappings using 3 arrays API
by Richard W.M. Jones 09 Apr '08

09 Apr '08

Previous discussion: http://www.redhat.com/archives/libvir-list/2008-February/thread.html#00281 http://www.redhat.com/archives/libvir-list/2008-February/thread.html#00348 This patch also contains the fix to close the file descriptor mentioned by Jim Meyering in that second thread. Hugh: Is this still wanted for oVirt? If oVirt don't specifically need it then we may as well leave it out. Rich. -- Richard Jones, Emerging Technologies, Red Hat http://et.redhat.com/~rjones virt-p2v converts physical machines to virtual machines. Boot with a live CD or over the network (PXE) and turn machines into Xen guests. http://et.redhat.com/~rjones/virt-p2v

3 2

[Libvir] [PATCH] Core code contains Xen-ism UUID of dom0 = 0000-0000-0000-0000
by Richard W.M. Jones 09 Apr '08

09 Apr '08

This patch removes a Xen-ism from the core code. The Xen driver sets the UUID to all-zeroes so we don't need a hack in virDomainGetUUID. Rich. -- Richard Jones, Emerging Technologies, Red Hat http://et.redhat.com/~rjones virt-top is 'top' for virtual machines. Tiny program with many powerful monitoring features, net stats, disk stats, logging, etc. http://et.redhat.com/~rjones/virt-top

4 3

[Libvir] [PATCH REPOST] warn about conn, dom and net fields in virterror
by Richard W.M. Jones 09 Apr '08

09 Apr '08

First discussed in this thread: http://www.redhat.com/archives/libvir-list/2008-March/thread.html#00367 The attached patch contains a slightly different warning about the danger of using these fields. Rich. -- Richard Jones, Emerging Technologies, Red Hat http://et.redhat.com/~rjones virt-top is 'top' for virtual machines. Tiny program with many powerful monitoring features, net stats, disk stats, logging, etc. http://et.redhat.com/~rjones/virt-top

4 3

[Libvir] [PATCH] Add virsh.pod to EXTRA_DIST
by Richard W.M. Jones 09 Apr '08

09 Apr '08

-- Richard Jones, Emerging Technologies, Red Hat http://et.redhat.com/~rjones virt-p2v converts physical machines to virtual machines. Boot with a live CD or over the network (PXE) and turn machines into Xen guests. http://et.redhat.com/~rjones/virt-p2v

2 1

[libvir] [PATCH] Extra symbol in openvz_conf.c
by Anton Protopopov 09 Apr '08

09 Apr '08

Hi all, there is an extra symbol ('_') in openvz_conf.c: openvzParseXML, which leads to compile error [does anybody compile libvirt with openvz support :)] Patch fixes it.

3 2

[Libvir] Release of libvirt-0.4.2
by Daniel Veillard 09 Apr '08

09 Apr '08

A lot of things accumulated since last month, especially bug fixes, so it was time for a new release, available at ftp://libvirt.org/libvirt/ I didn't list Linux Container support in the new features as this isn't complete yet (but looks like it will be shortly) * New features: - memory operation for QEmu/KVM driver (Cole Robinson) - new routed networking schemas (Mads Olesen) * Documentation: - storage documentation fixes (Atsushi Sakai) - many typo cleanups (Atsushi Sakai) - string fixes (Francesco Tombolini) * Bug fixes: - pointer errors in qemu (Jim Meyering) - iSCSI login fix (Chris Lalancette) - well formedness error in test driver capabilities (Cole Robinson) - fixes cleanup code when daemon exits (Daniel Berrange) - CD Rom change on live QEmu/KVM domains (Cole Robinson) - setting scheduler parameter is forbidden for read-only (Saori Fukuta)i - fixes for TAP devices (Daniel Berrange) - assorted storage driver fixes (Daniel Berrange) - Makefile fixes (Jim Meyering) - Xen-3.2 hypercall fix - fix iptables rules to avoid blocking traffic within virtual network (Daniel Berrange) - XML output fix for directory pools (Daniel Berrange) - remove dandling domain/net/conn pointers from error data - do not ask polkit auth when root (Daniel Berrange) - handling of fork and pipe errors when starting the daemon (Richard Jones) * Improvements: - better validation of MAC addresses (Jim Meyering and Hiroyuki Kaguchi) - virsh vcpupin error report (Shigeki Sakamoto) - keep boot tag on HVM domains (Cole Robinson) - virsh non-root should not be limited to read only anymore (Daniel Berrange) - switch to polkit-auth from polkit-grant (Daniel Berrange) - better handling of missing SElinux data (Daniel Berrange and Jim Meyering) - cleanup of the connection opening logic (Daniel Berrange) - first bits of Linux Containers support (Dave Leskovec) - scheduler API support via xend (Saori Fukuta) - improvement of the testing framework and first tests (Jim Meyering) - missing error messages from virsh parameters validation (Shigeki Sakamoto) - improve support of older iscsiadm command (Chris Lalancette) - move linux container support in the daemon (Dan Berrange) - older awk implementation support (Mike Gerdts) - NUMA support in test driver (Cole Robinson) - xen and hvm added to test driver capabilities (Cole Robinson) * Code cleanup: - remove unused getopt header (Jim Meyering) - mark more strings as translatable (Guido Günther and Jim Meyering) - convert error strings to something meaningful and translatable (Jim Meyering) - Linux Containers code cleanup - last error initializer (Guido Günther) Quite a lot of changes in a month, thanks a lot for everybody who contributed with ideas, reports or patches ! Daniel -- Red Hat Virtualization group http://redhat.com/virtualization/ Daniel Veillard | virtualization library http://libvirt.org/ veillard(a)redhat.com | libxml GNOME XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/

4 5

[Libvir] [PATCH] (for discussion only) KVM migration v1
by Richard W.M. Jones 09 Apr '08

09 Apr '08

This patch implements KVM to KVM migration. It is for discussion only, partly because it doesn't work for some reason I can't quite work out at the moment. We implement a second version of the migration protocol. This second version has two differences: (1) Prepare step is replaced by Prepare2, which passes the domain XML. As explained previously this is required for KVM because we use this on the target host to recreate the precise qemu-kvm command line as on the source host. (2) Finish step is replaced by Finish2. There are two differences here: firstly Finish2 is always called on the destination host, even if the migration failed. Secondly the return code from the migration (Perform step) is passed to Finish2. This is required for KVM migration because if the migration failed we need to tear down the empty qemu shell, otherwise failed migrations could leave effectively zombie qemu processes around. libvirt.c:virDomainMigrate function has been changed so that it can support either form of migration protocol, and the Xen driver continues to use version 1. The changes here are pretty minor, and there are no changes to the Xen driver. The additional code involves implementing Prepare2 / Finish2 in the remote protocol, and of course implementing migration in the qemu driver itself. A final word about the parameters to virDomainMigrate. - flags is ignored. All KVM migrations are "live", it doesn't matter if you supply the live flag or not. - desturi may optionally be used to control the TCP port used for migration. If desturi is NULL then a TCP port is chosen at random (or in future, some suitable secure method will be used instead). If desturi is set to "tcp://hostname:port" then the given port number is used, and hostname is expected to be the hostname or IP address of the target server. You cannot do localhost->localhost migrations (even though this is supported by KVM) because libvirtd doesn't like you creating two VMs with the same UUID, even if only temporarily. So to test this you really need two machines, or at least two instances of libvirtd configured not to stomp on each other. Rich. -- Richard Jones, Emerging Technologies, Red Hat http://et.redhat.com/~rjones virt-top is 'top' for virtual machines. Tiny program with many powerful monitoring features, net stats, disk stats, logging, etc. http://et.redhat.com/~rjones/virt-top

2 1

[Libvir] <model>e1000</model> - Specifying model in XML Configs
by Henri Cook 08 Apr '08

08 Apr '08

Hi guys, I just tried to port a few customers over to this libvirt setup i'm hoping to run and I couldn't get their machines started because there's no mechanism to specify a model in the XML! I know redhat have changed KVM's default driver to e1000, which I think is the one I like the most and almost if not all of my VMs use. Ubuntu haven't done that yet however! Is there a model directive in the latest version? In the works? Just so I know which solution I should pursue (updating libvirt or harassing ubuntu to change the default) Thanks, Henri

5 10

[Libvir] Don't fail to read a file because it's non-seekable (e.g., a pipe).
by Jim Meyering 08 Apr '08

08 Apr '08

FYI, I expect to add fread_file_lim or something like it to gnulib, and it already has some unit tests (passed). I removed the "* tab-width: 4" line because it seriously mangled the code that I initially added. Besides saying "tab-width 4" is contradictory with the "indent-tabs-mode: nil" setting. This fix addresses a problem exposed in an ovirt script whereby trying to use bash process substitution, e.g., in virsh define <(command to generate xml) would fail. Oops. Just noticed that the indentation in the added function (gnulib style) is not consistent with the rest of the file. I'll adjust that before committing, of course. Don't fail to read a file because it's non-seekable (e.g., a pipe). * src/util.c (fread_file_lim): New function. (__virFileReadAll): Use fread_file_lim, rather than requiring that stat.st_size provide a usable file size. * tests/read-non-seekable: New test, for the above. * tests/Makefile.am (test_scripts): Add read-non-seekable. * tests/test-lib.sh (mkfifo_or_skip_): New helper function. Signed-off-by: Jim Meyering <meyering(a)redhat.com> --- src/util.c | 87 +++++++++++++++++++++++++++++++++++++--------- tests/Makefile.am | 1 + tests/read-non-seekable | 47 +++++++++++++++++++++++++ tests/test-lib.sh | 12 ++++++ 4 files changed, 130 insertions(+), 17 deletions(-) create mode 100755 tests/read-non-seekable diff --git a/src/util.c b/src/util.c index 0667780..e951eb5 100644 --- a/src/util.c +++ b/src/util.c @@ -49,6 +49,10 @@ #include "util-lib.c" +#ifndef MIN +# define MIN(a, b) ((a) < (b) ? (a) : (b)) +#endif + #define MAX_ERROR_LEN 1024 #define TOLOWER(Ch) (isupper (Ch) ? tolower (Ch) : (Ch)) @@ -283,6 +287,61 @@ virExecNonBlock(virConnectPtr conn, #endif /* __MINGW32__ */ +/* Like gnulib's fread_file, but read no more than the specified maximum + number of bytes. If the length of the input is <= max_len, and + upon error while reading that data, it works just like fread_file. */ +static char * +fread_file_lim (FILE *stream, size_t max_len, size_t *length) +{ + char *buf = NULL; + size_t alloc = 0; + size_t size = 0; + int save_errno; + + for (;;) + { + size_t count; + size_t requested; + + if (size + BUFSIZ + 1 > alloc) + { + char *new_buf; + + alloc += alloc / 2; + if (alloc < size + BUFSIZ + 1) + alloc = size + BUFSIZ + 1; + + new_buf = realloc (buf, alloc); + if (!new_buf) + { + save_errno = errno; + break; + } + + buf = new_buf; + } + + /* Ensure that (size + requested <= max_len); */ + requested = MIN (size < max_len ? max_len - size : 0, + alloc - size - 1); + count = fread (buf + size, 1, requested, stream); + size += count; + + if (count != requested || requested == 0) + { + save_errno = errno; + if (ferror (stream)) + break; + buf[size] = '\0'; + *length = size; + return buf; + } + } + + free (buf); + errno = save_errno; + return NULL; +} int __virFileReadAll(const char *path, int maxlen, @@ -291,6 +350,8 @@ int __virFileReadAll(const char *path, FILE *fh; struct stat st; int ret = -1; + size_t len; + char *s; if (!(fh = fopen(path, "r"))) { virLog("Failed to open file '%s': %s", @@ -309,28 +370,21 @@ int __virFileReadAll(const char *path, goto error; } - if (st.st_size > maxlen) { - virLog("File '%s' is too large %d, max %d", path, (int)st.st_size, maxlen); - goto error; - } - - *buf = malloc(st.st_size + 1); - if (*buf == NULL) { - virLog("Failed to allocate data"); + s = fread_file_lim(fh, maxlen+1, &len); + if (s == NULL) { + virLog("Failed to read '%s': %s", path, strerror (errno)); goto error; } - if ((ret = fread(*buf, st.st_size, 1, fh)) != 1) { - free(*buf); - *buf = NULL; - virLog("Failed to read config file '%s': %s", - path, strerror(errno)); + if (len > maxlen || (int)len != len) { + free(s); + virLog("File '%s' is too large %d, max %d", + path, (int)st.st_size, maxlen); goto error; } - (*buf)[st.st_size] = '\0'; - - ret = st.st_size; + *buf = s; + ret = len; error: if (fh) @@ -739,6 +793,5 @@ virParseMacAddr(const char* str, unsigned char *addr) * indent-tabs-mode: nil * c-indent-level: 4 * c-basic-offset: 4 - * tab-width: 4 * End: */ diff --git a/tests/Makefile.am b/tests/Makefile.am index 901e88a..ca12b84 100644 --- a/tests/Makefile.am +++ b/tests/Makefile.am @@ -46,6 +46,7 @@ noinst_PROGRAMS = xmlrpctest xml2sexprtest sexpr2xmltest virshtest conftest \ test_scripts = \ daemon-conf \ int-overflow \ + read-non-seekable \ vcpupin EXTRA_DIST += $(test_scripts) diff --git a/tests/read-non-seekable b/tests/read-non-seekable new file mode 100755 index 0000000..9bb4a21 --- /dev/null +++ b/tests/read-non-seekable @@ -0,0 +1,47 @@ +#!/bin/sh +# ensure that certain file-reading commands can handle non-seekable files + +# Copyright (C) 2008 Free Software Foundation, Inc. + +# This program is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation, either version 3 of the License, or +# (at your option) any later version. + +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. + +# You should have received a copy of the GNU General Public License +# along with this program. If not, see <http://www.gnu.org/licenses/>. + +if test "$VERBOSE" = yes; then + set -x + virsh --version +fi + +. $srcdir/test-lib.sh + +fail=0 + +cat <<\EOF > dom +<domain type='test' id='2'> + <name>t2</name> + <uuid>004b96e1-2d78-c30f-5aa5-000000000000</uuid> + <memory>8388608</memory> + <vcpu>2</vcpu> + <on_reboot>restart</on_reboot> + <on_poweroff>destroy</on_poweroff> + <on_crash>restart</on_crash> +</domain> +EOF + +virsh -c test:///default define dom > /dev/null || fail=1 + +mkfifo_or_skip_ fifo +cat dom > fifo & + +virsh -c test:///default define fifo > /dev/null || fail=1 + +(exit $fail); exit $fail diff --git a/tests/test-lib.sh b/tests/test-lib.sh index cdbea5d..a007109 100644 --- a/tests/test-lib.sh +++ b/tests/test-lib.sh @@ -120,6 +120,18 @@ skip_if_root_() { uid_is_privileged_ && skip_test_ "must be run as non-root"; } error_() { echo "$0: $@" 1>&2; (exit 1); exit 1; } framework_failure() { error_ 'failure in testing framework'; } +mkfifo_or_skip_() +{ + test $# = 1 || framework_failure + if ! mkfifo "$1"; then + # Make an exception of this case -- usually we interpret framework-creation + # failure as a test failure. However, in this case, when running on a SunOS + # system using a disk NFS mounted from OpenBSD, the above fails like this: + # mkfifo: cannot make fifo `fifo-10558': Not owner + skip_test_ 'NOTICE: unable to create test prerequisites' + fi +} + test_dir_=$(pwd) this_test_() { echo "./$0" | sed 's,.*/,,'; } -- 1.5.5.rc2.7.g144a

3 6