- Devel - libvirt List Archives

[libvirt] [patch][doc] fix typos 20080807
by Atsushi SAKAI 07 Aug '08

07 Aug '08

Hi, Typo fixes on document. formatdomain.html | 22 +++++++++++----------- formatdomain.html.in | 22 +++++++++++----------- java.html | 2 +- java.html.in | 2 +- 4 files changed, 24 insertions(+), 24 deletions(-) Thanks Atsushi SAKAI

2 1

[libvirt] How does virt-manager send shutdown?
by Bryan Kearney 07 Aug '08

07 Aug '08

How does virt-manager send shutdown to guests? Does it depend on if you are running kvm or xen? I have been building images with acpid enabled and acpi=force on the command line. After upgrading to f9 and running kvm my images no longer respond to shutdown. Is there a tweak which I have missed? - bk

5 8

[libvirt] [PATCH] storage pool discovery
by David Lively 06 Aug '08

06 Aug '08

Hi - The attached patch implements virConnectDiscoverStoragePools mostly as specified by Daniel Berrange in: http://www.redhat.com/archives/libvir-list/2008-February/msg00107.html Daniel wasn't happy with the interface for this function because it wasn't sufficiently general, so it currently doesn't exist. I've attempted to generalize his proposed interface by changing the "hostname" argument to a "srcSpec" argument, which is intended to be a XML storage pool <source> element. Note that srcSpec is not necessary for some (local) pool types. For example, to discover existing LVM volume groups, there's no need to specify a source: virsh # pool-discover logical While discovering nfs shares requires a <source> document to specify the host to query: virsh # pool-discover netfs "<source><host name='share' /></source>" Currently (i.e., in this patch) I've implemented discovery support only for the "logical" and "netfs" pools. I plan on covering the other types once we've agreed upon the API. While this patch is rather large, the vast majority of it is "generic" (not specific to any particular type of storage pool) code, which I mostly took from Daniel's submission. I could easily break the storage-pool-type-specific pieces into separate patches, but those changes are already segregated into the appropriate storage_backend_ files. * Known Issues * I made the virsh interface take a XML string rather than a filename. I found it convenient, but AFAIK the rest of the commands that take XML take it from a file, so perhaps that should be changed (just let me know what you'd prefer). I realize the description of srcSpec is kind of fuzzy. For instance, for netfs discovery, the <source> must have a <host> element (to specify the host to query), but shouldn't have a <dir> element since the exported dirs are exactly what we're trying to discover in this case. So this really needs to be documented accurately for each storage pool type. (Where?) Finally, there's an underspecification issue. The XML pool descriptions returned are supposed to be usable as valid input to virStoragePoolDefineXML. But these descriptions include some data (pool name, target path) that isn't specified by the discover input or the discovered resources. For now, I'm making up a somewhat arbitrary pool name (logical: VolGroup name, netfs: last component of export path). And I don't even specify <target><path> in the "netfs" discovery output (which means it's not valid input to virStoragePoolDefineXML until a target path is added). Dave

4 14

[libvirt] [PATCH] fix failing "make syntax-check"
by Jim Meyering 06 Aug '08

06 Aug '08

"make syntax-check" just started failing. This fixes it: fix failing "make syntax-check" * src/stats_linux.c: Don't include <c-ctype.h> no longer used. --- ChangeLog | 3 +++ src/stats_linux.c | 1 - 2 files changed, 3 insertions(+), 1 deletions(-) diff --git a/ChangeLog b/ChangeLog index c23799c..ea6f1da 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,5 +1,8 @@ Tue Aug 6 13:30:44 CEST 2008 Jim Meyering <meyering(a)redhat.com> + fix failing "make syntax-check" + * src/stats_linux.c: Don't include <c-ctype.h> no longer used. + work around MinGW build failure due to its lack of inet_aton Use inet_pton instead; pull in gnulib's module by the same name. * src/network_conf.c (virNetworkDefParseXML): Use inet_pton, diff --git a/src/stats_linux.c b/src/stats_linux.c index 897251b..3f7dd52 100644 --- a/src/stats_linux.c +++ b/src/stats_linux.c @@ -19,7 +19,6 @@ #include <string.h> #include <unistd.h> #include <regex.h> -#include "c-ctype.h" #ifdef WITH_XEN #include <xs.h> -- 1.6.0.rc1.92.g189f7

2 2

[libvirt] Memory corruption with CVS tip + python
by Chris Lalancette 06 Aug '08

06 Aug '08

All, FYI, I'm running into some memory corruption when using CVS tip and trying to use an older virt-install command. I won't have time to debug this today, so I thought I would mention it on-list in case someone else wants to take a crack at it. My system is an AMD system running RHEL-5.2 x86_64 bits, so it has python-virtinst-0.300.2-8.el5 installed. Basically, I did this: # cvs co libvirt # cd libvirt # ./autogen.sh # make dist >From here, I took the resulting .tar.gz, and dumped it into /usr/src/redhat/SOURCES, and then took the SPEC file from the Fedora 9 sources. I had to modify the very top of it (since RHEL-5 rpmbuild didn't like the macros for whatever reason), but if I read the logic right, I put these flags at the top: %define with_polkit 0 %define with_proxy yes %define with_qemu 0 Then I built and installed a libvirt and libvirt-python RPM with this. Finally, I restarted libvirtd, and then tried to run the following command: [root@amd1 libvirt]# virt-install -n test -f /var/lib/xen/images/test.img -s 10 -r 768 -v -c /var/lib/xen/images/boot.iso -d --vnc Wed, 06 Aug 2008 07:59:10 DEBUG DistroInstaller location is a local file/path: /var/lib/xen/images/boot.iso Starting install... Segmentation fault [root@amd1 libvirt]# It happens every time, so it should be fairly easy to track down. Chris Lalancette

1 0

[libvirt] incorrect data in online documentation (I think)
by matthew chan 06 Aug '08

06 Aug '08

Hi The online documentation mentions that the XML <memory> node units are in byte, but from the example code I have observed, it seems more likely to be in kilobytes. Could someone verify + change if necessary please? Thanks, Matt

2 1

[libvirt] [PATCH]: Fix sexpr generation again
by Chris Lalancette 06 Aug '08

06 Aug '08

DanB pointed out that my sexpr xend_internal patch from yesterday broke one of the regression tests. The problem is that the xenDaemonFormatSxpr{Disk,Net} functions are shared between domain creation time and attaching disk time. Unfortunately, though, Xend expects something different during these two times. During domain creation time, it wants the "(device" in front of the sexpr, while during attach time it does not. To remedy this situation, I added a flag to these two functions to differentiate between these two modes. With this patch in place, all of the regression tests pass. Signed-off-by: Chris Lalancette <clalance(a)redhat.com>

3 2

Re: [libvirt] PATCH: 0/7: Re-factor LXC driver
by Daniel P. Berrange 05 Aug '08

05 Aug '08

On Tue, Aug 05, 2008 at 09:04:34PM +0400, Evgeniy V. Sokolov wrote: > > >This is a long overdue followup to my previous set of patches to make the > >LXC driver use the new domain XML apis. Getting this merged is a blocker > >for the new libvirt release, now we've turned on LXC by default. Likewise > >we need to get OpenVZ fully ported to new XML before release too. > Daniel, do you plan to finish port of OpenVZ ? or I will do it. I'm happy todo the work to at least port the code, but I don't have a OpenVZ enabled machine, so would neeed someone else to actually test if its works :-) Ideally I think we should aim todo a release next week, so I'll look at whatever the state of the OpenVZ code is this coming weekend, and work from that. Regards, Daniel -- |: Red Hat, Engineering, London -o- http://people.redhat.com/berrange/ :| |: http://libvirt.org -o- http://virt-manager.org -o- http://ovirt.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: GnuPG: 7D3B9505 -o- F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :|

2 1

[libvirt] [PATCH]: Fix "virsh attach-disk" and "virsh attach-interface"
by Chris Lalancette 05 Aug '08

05 Aug '08

With the recent refactoring of the domain code, plus the changes with the Xend code, a couple of bugs were introduced into the attach-disk and attach-interface functionality. This patch fixes 3 bugs: 1) In xenDaemonAttachDevice(), there is a switch statement to determine which of the xenDaemonFormatSxpr{Disk,Net} functions to call. Unfortunately, the case statements are all missing the corresponding "break", so we always fall-through to the default error case. This patch just adds the appropriate break statements. 2) (minor) In xenDaemonDomainDefineXML (that's a mouthful!), there is a stray "fprintf". This is now converted to a proper virXendError(). 3) xenDaemonFormatSxpr{Disk,Net} were adding an extra (device to the front of the sexpr expressions that xend did not expect (this is Xend on RHEL 5.2). Because of this, the attaches would fail. The patch fixes this by removing the (device from the front, which makes attach-disk and attach-interface work again. Signed-off-by: Chris Lalancette <clalance(a)redhat.com>

3 4

[libvirt] [PATCH]: Rework xenLinuxDomainDeviceID to support more devices
by Chris Lalancette 05 Aug '08

05 Aug '08

Recently upstream Xen added support for having xvd devices > 16. For the most part, this doesn't really concern libvirt, since for things like attach and detach we just pass it through and let xend worry about whether it is supported or not. The one place this breaks down is in the stats collecting code, where we need to figure out the device number so we can go digging in /sys for the statistics. To remedy this, I've re-written xenLinuxDomainDeviceID() to use regular expressions to figure out the device number from the name. The major advantage is that now xenLinuxDomainDeviceID() looks fairly identical to tools/python/xen/util/blkif.py (in the Xen sources), so that adding additional devices in the future should be much easier. It also reduces the size of the code, and, in my opinion, the code complexity. With this patch in place, I was able to get block statistics both on older style devices (/dev/xvda) and on the new, expanded devices (/dev/xvdaa). Signed-off-by: Chris Lalancette <clalance(a)redhat.com>

4 10