- Devel - libvirt List Archives

[libvirt] Is CVS server working?
by Atsushi SAKAI 07 Aug '08

07 Aug '08

Hi, CVS server anoncvs@libvirt.org:2401/data/cvs is working? I can see git repository but CVS repository are not. Thanks Atsushi SAKAI

2 3

[libvirt] [PATCH] Install default storage pool
by Cole Robinson 07 Aug '08

07 Aug '08

The attached patch adds a default pool definition to install with libvirt. The pool is a directory pool, hardcoded to use /var/lib/libvirt/images, though this is replaced with %{_localstatedir}/lib/libvirt/images if installing via rpm. Since /var/lib/libvirt/images may not exist if installing from tar.gz, the pool is set not to autostart by default via that method. This patch isn't ready to apply though, since there is one issue I don't know how to solve. We don't want every rpm upgrade to install the default pool, since we don't want the default to continually resurrect itself if a user explicitly deletes it. The default network handles this by checking if libvirtd has been installed. Since libvirtd and virtual network support were introduced around the same time, this makes sense. However for storage, I can't think of any simple check we can make. If we used the same as the default network, all f8 or f9 upgrades wouldn't install the default pool. I can't really think of anything to check, besides maybe doing some sort of time comparison on when /usr/share/libvirt/storage/default.xml was accessed. This code is just hardcoded to always add it. Any ideas? Thanks, Cole

2 1

[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