2011/2/15 Osier Yang <jyang@redhat.com>:

于 2011年02月15日 11:32, Daniel Veillard 写道:

...

  Thanks everybody for the testing feedback and fixes, in retrospect I should really had done this in previous releases ! So the third rc tarball is out, it's likely to be the last one before the release (within 48 hours):

   ftp://libvirt.org/libvirt/libvirt-0.8.8-rc3.tar.gz

  give it a try !

   thanks,

Daniel

# ./daemon/libvirtd 14:11:15.018: 31796: info : libvirt version: 0.8.8, package:  ()

...

...

...

PACKAGE_VERSION and PACKAGER are all defined as "" by default, do we need to do more checking to print just the following?

14:11:15.018: 31796: info : libvirt version: 0.8.8

14:11:15.018: 31796: error : virCommandWait:1226 : internal error Child process exited with status 2. 14:11:15.038: 31796: error : virCommandWait:1226 : internal error Child process exited with status 2.

...

...

...

Not sure if above two errors are allowed, didn't investigate.

Regards Osier

I have this two errors two. The come from the network driver trying to use iptables features that are not present in my iptables version. The driver expects this and handles it by outputting a warning in that case: 11:02:01.686: 13937: warning : networkAddGeneralIptablesRules:1128 : Could not add rule to fixup DHCP response checksums on network 'default'. 11:02:01.686: 13937: warning : networkAddGeneralIptablesRules:1129 : May need to update iptables package & kernel to support CHECKSUM rule. But the virCommand code reports the failing call to iptables as error. We might want to add some logic to the virCommand code to suppress expected errors. You can test if it's the same problem for you by applying this patch: diff --git a/src/util/command.c b/src/util/command.c index abd2dc4..258dc13 100644 --- a/src/util/command.c +++ b/src/util/command.c @@ -1221,9 +1221,11 @@ virCommandWait(virCommandPtr cmd, int *exitstatus) if (exitstatus == NULL) { if (status != 0) { + char *str = virCommandToString(cmd); virCommandError(VIR_ERR_INTERNAL_ERROR, - _("Child process exited with status %d."), - WEXITSTATUS(status)); + _("Child process (%s) exited with status %d."), + str ? str : cmd->args[0], WEXITSTATUS(status)); + VIR_FREE(str); return -1; } } else { that makes it output the failing command. Matthias

2011/2/15 Osier Yang <jyang@redhat.com>:

于 2011年02月15日 11:32, Daniel Veillard 写道:

...

  Thanks everybody for the testing feedback and fixes, in retrospect I should really had done this in previous releases ! So the third rc tarball is out, it's likely to be the last one before the release (within 48 hours):

   ftp://libvirt.org/libvirt/libvirt-0.8.8-rc3.tar.gz

  give it a try !

   thanks,

Daniel

# ./daemon/libvirtd 14:11:15.018: 31796: info : libvirt version: 0.8.8, package:  ()

...

...

...

PACKAGE_VERSION and PACKAGER are all defined as "" by default, do we need to do more checking to print just the following?

configure tests --with-packager(-version) for "no" instead of empty string. I think --with-packager(-version) should either default to "no" diff --git a/configure.ac b/configure.ac index 3cd824a..9caa902 100644 --- a/configure.ac +++ b/configure.ac @@ -32,11 +32,11 @@ AC_SUBST([LIBVIRT_VERSION_NUMBER]) AC_ARG_WITH([packager], [AS_HELP_STRING([--with-packager], [Extra packager name])], - [],[]) + [],[with_packager=no]) AC_ARG_WITH([packager-version], [AS_HELP_STRING([--with-packager-version], [Extra packager version])], - [],[]) + [],[with_packager_version=no]) if test "x$with_packager" != "xno" then AC_DEFINE_UNQUOTED([PACKAGER], ["$with_packager"], or be tested for empty string diff --git a/configure.ac b/configure.ac index 3cd824a..2f371a7 100644 --- a/configure.ac +++ b/configure.ac @@ -37,12 +37,12 @@ AC_ARG_WITH([packager-version], [AS_HELP_STRING([--with-packager-version], [Extra packager version])], [],[]) -if test "x$with_packager" != "xno" +if test "x$with_packager" != "x" then AC_DEFINE_UNQUOTED([PACKAGER], ["$with_packager"], [Extra package name]) fi -if test "x$with_packager_version" != "xno" +if test "x$with_packager_version" != "x" then AC_DEFINE_UNQUOTED([PACKAGER_VERSION], ["$with_packager_version"], [Extra package version]) both approaches fix the cosmetic problem. Matthias

On Wed, Feb 16, 2011 at 05:10:10PM -0700, Jim Fehlig wrote:

Jim Fehlig wrote:

...

Daniel Veillard wrote:

...

Thanks everybody for the testing feedback and fixes, in retrospect I should really had done this in previous releases ! So the third rc tarball is out, it's likely to be the last one before the release (within 48 hours):

ftp://libvirt.org/libvirt/libvirt-0.8.8-rc3.tar.gz

give it a try !

Building and passing minimal tests on all supported openSUSE/SLE distros, which in itself is a good exercise of the various configure options.

However, I'm trying to run libvirt-tck on openSUSE11.4rc1, qemu0.14rc1, and libvirt 0.8.8rc3 (lots of rc's :)) and getting quite a few failures. The failing qemu invocations look like:

2011-02-15 12:15:05.863: starting up LC_ALL=C PATH=/sbin:/usr/sbin:/usr/local/sbin:/root/bin:/usr/local/bin:/usr/bin:/bin:/usr/bin/X11:/usr/X11R6/bin:/usr/games:/usr/lib64/jvm/jre/bin HOME=/root USER=root LOGNAME=root /usr/bin/qemu-kvm -S -M pc-0.14 -enable-kvm -m 64 -smp 1,sockets=1,cores=1,threads=1 -name tck -uuid 7767756e-9160-41d3-f7a5-0af2aed17d1b -nographic -nodefconfig -nodefaults -chardev socket,id=charmonitor,path=/var/lib/libvirt/qemu/tck.monitor,server,nowait -mon chardev=charmonitor,id=monitor,mode=control -rtc base=utc -boot c -kernel /var/cache/libvirt-tck/os-i686-hvm/vmlinuz -initrd /var/cache/libvirt-tck/os-i686-hvm/initrd -drive file=/var/cache/libvirt-tck/os-i686-hvm/disk.img,if=none,id=drive-virtio-disk0,boot=on,format=raw -device virtio-blk-pci,bus=pci.0,addr=0x2,drive=drive-virtio-disk0,id=virtio-disk0 -chardev pty,id=charserial0 -device isa-serial,chardev=charserial0,id=serial0 -usb -device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x3 char device redirected to /dev/pts/1 Two devices with same boot index 0

I think qemu0.14 is less tolerable of specifying 2 boot devices, in this case implicitly with -kernel and boot=on for the disk.

I'll try to look more later but for now need some sleep.

I see Bruce brought this up on qemu ml [1]. Seems removing boot=on when using linuxboot.bin is preferred solution. I'm attaching a hack that allows me to pass the libvirt-tck domain tests but would really like comment from someone familiar with code, particularly the bootindex stuff in qemu-kvm.

Thanks! Jim

[1] http://lists.nongnu.org/archive/html/qemu-devel/2011-02/msg01892.html

diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index 05f427c..9186707 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -3112,7 +3112,7 @@ qemuBuildCommandLine(virConnectPtr conn, }

/* If QEMU supports -drive param instead of old -hda, -hdb, -cdrom .. */ - if (qemuCmdFlags & QEMUD_CMD_FLAG_DRIVE) { + if (qemuCmdFlags & QEMUD_CMD_FLAG_DRIVE && !def->os.kernel) { int bootCD = 0, bootFloppy = 0, bootDisk = 0;

/* If QEMU supports boot=on for -drive param... */

This patch prevents libvirt actually creating any disks at all. Your idea sounds good though. I think if you put the '&& !def->os.kernel' bit in the next conditional it will work, eg try /* If QEMU supports boot=on for -drive param... */ if ((qemuCmdFlags & QEMUD_CMD_FLAG_DRIVE_BOOT) && !def->os.kernel) { Regards, Daniel -- |: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|