[libvirt] RHEL guest shuts down after running anaconda instead of rebooting
by Eric
when i create a RHEL 6 guest using virt-install it will go through the
process of installing the OS and want to reboot once finished. when it
reboots the VM powers off and i have to start it manually. is there a way to
prevent this? i can reboot guests once they're up and running and they'll
come back up without issue, but they always shut down when i'm installing
them. host OS is RHEL 6 x86_64.
13 years, 4 months
[libvirt] CentOS/RHEL 6 RPMs repository
by Mohammed Naser
Hi list!
First post here, thanks for a wonderful application, I have been around in
#virt on IRC and just signed up today. My little contribution:
We use CentOS 6 in production and we were not able to get access to the
latest 0.9.3 RPMs and we needed some of the new features, so we compiled a
bunch of RPMs to help those save time (and us). Please note that we're not
responsible if it blows up your computer, etc. etc.
wget http://repo.vexxhost.com/virt/virt.repo -O /etc/yum.repos.d/virt.repo
yum install libvirt
That should get you started on libvirt 0.9.3, there are a couple of
unavailable dependencies that are also on the repo, so you should be good to
go!
If anyone has any issues, please let me know. Other than that, enjoy the
repos! Also, if you don't trust us and want to compile the SRPMS yourself,
they are also available -- if you are running i386 and compile a set, drop
it by so we can help people as it is x86_64 RPMs only right now.
Regards,
--
Mohammed Naser — vexxhost
-----------------------------------------------------
13 years, 4 months
[libvirt] [PATCH] maint: fix spelling of Red Hat
by Eric Blake
* cfg.mk (sc_copyright_format): Add to rule.
* src/util/files.h: Fix offenders.
* src/util/files.c: Likewise.
---
Pushing under the trivial rule. Noticed it because my upcoming edits to
files.c triggered my editor's copyright update rule. The 'syntax-check'
addition ensures that no one else makes the mistake.
cfg.mk | 3 +++
src/util/files.c | 4 ++--
src/util/files.h | 2 +-
3 files changed, 6 insertions(+), 3 deletions(-)
diff --git a/cfg.mk b/cfg.mk
index f98c94b..d243862 100644
--- a/cfg.mk
+++ b/cfg.mk
@@ -584,6 +584,9 @@ sc_copyright_format:
@prohibit='Copyright [^(].*Red 'Hat \
halt='consistently use (C) in Red Hat copyright' \
$(_sc_search_regexp)
+ @prohibit='\<Red''Hat\>' \
+ halt='spell Red Hat as two words' \
+ $(_sc_search_regexp)
# Some functions/macros produce messages intended solely for developers
# and maintainers. Do not mark them for translation.
diff --git a/src/util/files.c b/src/util/files.c
index de4ae87..bef56b6 100644
--- a/src/util/files.c
+++ b/src/util/files.c
@@ -1,9 +1,9 @@
/*
- * memory.c: safer file handling
+ * files.c: safer file handling
*
+ * Copyright (C) 2010-2011 Red Hat, Inc.
* Copyright (C) 2010 IBM Corporation
* Copyright (C) 2010 Stefan Berger
- * Copyright (C) 2010 RedHat, Inc.
* Copyright (C) 2010 Eric Blake
*
* This library is free software; you can redistribute it and/or
diff --git a/src/util/files.h b/src/util/files.h
index 5137fb3..8b681eb 100644
--- a/src/util/files.h
+++ b/src/util/files.h
@@ -1,7 +1,7 @@
/*
* files.h: safer file handling
*
- * Copyright (C) 2010-2011 RedHat, Inc.
+ * Copyright (C) 2010-2011 Red Hat, Inc.
* Copyright (C) 2010 IBM Corporation
* Copyright (C) 2010 Stefan Berger
* Copyright (C) 2010 Eric Blake
--
1.7.4.4
13 years, 4 months
[libvirt] [PATCH 0/8] Avoid filesystem cache pollution during virsh save
by Eric Blake
Saving a domain's state creates a large file, which risks polluting
the filesystem cache and slowing down a system. If a system has
a lot of domians simultaneously being saved (such as the libvirt-guests
init script doing managed saves), then this can cause noticeable slowdown
due to filesystem thrashing.
This patch series has been successfully tested to do 'virsh save
dom file --direct', with lsof(1) used to verify that O_DIRECT was
in use, and the resulting file was successfully used with 'virsh
restore file'.
Still to come - wire up O_DIRECT on the 'virsh restore' path (yes,
that means adding virDomainRestoreFlags - if only we had had the
foresight to use flags everywhere). Wire up qemu.conf to allow
the automatic use of --direct on automatic core dumps. Wire up
libvirt-guests init script to allow the use of --direct.
But I had enough in place to get the review started now.
Also, this series demonstrates some of the points I was
making about adding a new Flags API in this thread:
https://www.redhat.com/archives/libvir-list/2011-July/msg00762.html
This series requires and was tested on top of these (un-acked) patches:
https://www.redhat.com/archives/libvir-list/2011-July/msg00670.html
https://www.redhat.com/archives/libvir-list/2011-July/msg00675.html
Eric Blake (8):
save: document new public API
save: wire up remote protocol
save: wire up trivial saveFlags implementations
save: add --direct flag to virsh save operations
save: let iohelper handle inherited fd
save: let iohelper work on O_DIRECT fds
save: add virDirectFd wrapper type
save: support O_DIRECT during qemu saves
cfg.mk | 1 +
configure.ac | 6 +-
include/libvirt/libvirt.h.in | 14 +++
po/POTFILES.in | 1 +
src/Makefile.am | 1 +
src/driver.h | 6 ++
src/fdstream.c | 32 +++-----
src/libvirt.c | 94 +++++++++++++++++++++-
src/libvirt_private.syms | 6 ++
src/libvirt_public.syms | 5 +
src/libxl/libxl_driver.c | 17 ++++-
src/qemu/qemu_driver.c | 76 +++++++++++++-----
src/remote/remote_driver.c | 1 +
src/remote/remote_protocol.x | 11 +++-
src/remote_protocol-structs | 6 ++
src/test/test_driver.c | 20 ++++-
src/util/iohelper.c | 185 +++++++++++++++++++++++++++++++----------
src/util/virdirect.c | 149 +++++++++++++++++++++++++++++++++
src/util/virdirect.h | 37 +++++++++
src/vbox/vbox_tmpl.c | 19 ++++-
src/xen/xen_driver.c | 17 ++++-
tools/virsh.c | 12 +++-
tools/virsh.pod | 17 +++-
23 files changed, 632 insertions(+), 101 deletions(-)
create mode 100644 src/util/virdirect.c
create mode 100644 src/util/virdirect.h
--
1.7.4.4
13 years, 4 months
Re: [libvirt] mingw: test-poll pipe part fails
by Eric Blake
[adding libvirt]
On 06/04/2011 12:24 AM, Paolo Bonzini wrote:
> On Sat, Jun 4, 2011 at 00:37, Matthias Bolte
> <matthias.bolte(a)googlemail.com> wrote:
>> After testing a while and reading MSDN docs the problem seems to be
>> that MsgWaitForMultipleObjects doesn't work on pipes. It doesn't
>> actually wait but just returns immediately. Digging MSDN and googling
>> about this suggest that there is no simple solution to this.
>
> Yes, Windows pipes are that broken. :(
>
> Using socketpair is a possibly good idea, but I would do it on
> libvirtd only. I don't know exactly how libvirtd uses this pipe, but
> perhaps it can be changed to an eventfd-like abstraction that can be
> used with both Windows and Unix. This is similar to Eric's
> suggestion, but without the pipe at all. It would also be a
> libvirtd-specific suggestion.
I'm wondering if the problem here is that libvirt is trying to use the
pipe-to-self mechanism as a fundamental event loop idiom. That is, the
reason libvirt is calling poll is in order to minimize CPU until
something interesting happens, where interesting includes needing to
wake up a helper thread to do an action inside locks in response to the
receipt of a signal.
Maybe you are on to something, and replacing all uses of pipe() with
virPipeToSelf() (which uses pipe() for efficiency on Linux, but
socketpair() on mingw), would allow libvirt to continue to use the
pipe-to-self idiom while also using fds that can actually be poll'd on
mingw.
--
Eric Blake eblake(a)redhat.com +1-801-349-2682
Libvirt virtualization library http://libvirt.org
13 years, 4 months
[libvirt] enumerating disk devices
by Serge E. Hallyn
Hi,
Some people appear to have autostart VMs residing on slow storage. If
libvirtd starts too early, it'll simply fail trying to start those VMs.
It'd be nice to know when all the storage on which autostart containers
depend becomes available, so as to safely start libvirt.
The python script appended below is one approach. It could be called as
something like
for f in /etc/libvirt/qemu/autostart/* /etc/libvirt/lxc/autostart/*; do
for d in `libvirt_list_storage.py $f`; do
if [ ! -r $d ]; then
echo "storage not yet ready"
exit 1
fi
done
done
echo "starting libvirtd"
libvirtd
(Realistically I would have this running in an upstart job which is
'start on mounted' and, if all storage available, emits a
'libvirt-storage-up' event).
Are there better ways to go about this?
thanks,
-serge
#!/usr/bin/python
import sys
from xml.dom.minidom import parse
dom = parse(sys.argv[1])
def getText(nodelist):
rc = []
for node in nodelist:
if node.nodeType == node.TEXT_NODE:
rc.append(node.data)
return ''.join(rc)
def print_domname(dom):
#return
print "Domain %s\n" % getText(dom.childNodes)
def liststorages(dom):
print_domname(dom.getElementsByTagName("name")[0])
disks = dom.getElementsByTagName("disk")
list_disks(disks)
filesystems = dom.getElementsByTagName('filesystem')
list_filesystems(filesystems)
def list_filesystems(fslist):
for fs in fslist:
if not fs.hasAttribute('type'):
continue
if fs.getAttribute('type') != 'mount':
continue
dirs = fs.getElementsByTagName("source")
if len(dirs) == 0:
continue
if not dirs[0].getAttribute('dir'):
continue
print "dir: %s\n" % dirs[0].getAttribute('dir')
def list_disks(disks):
for disk in disks:
if not disk.hasAttribute('type'):
continue
if disk.getAttribute('type') == 'file':
list_disk_type_file(disk)
elif disk.getAttribute('type') == 'block':
list_disk_type_block(disk)
else:
# not supported
continue
def list_disk_type_file(disk):
disksrcs = disk.getElementsByTagName("source")
if len(disksrcs) == 0:
return
if not disksrcs[0].hasAttribute('file'):
return
print "file: %s\n" % disksrcs[0].getAttribute('file')
def list_disk_type_block(disk):
devsrscs = disk.getElementsByTagName("source")
if len(devsrscs) == 0:
return
if not devsrscs[0].hasAttribute('dev'):
return
print "dev: %s\n" % devsrscs[0].getAttribute('dev')
liststorages(dom)
13 years, 4 months
[libvirt] [PATCH 1/2] Add some basic sanity checking of certificates before use
by Daniel P. Berrange
If the libvirt daemon or libvirt client is configured with bogus
certificates, it is very unhelpful to only find out about this
when a TLS connection is actually attempted. Not least because
the error messages you get back for failures are incredibly
obscure.
This adds some basic sanity checking of certificates at the
time the virNetTLSContext object is created. This is at libvirt
startup, or when creating a virNetClient instance.
This checks that the certificate expiry/start dates are valid
and that the certificate is actually signed by the CA that is
loaded.
* src/rpc/virnettlscontext.c: Add certificate sanity checks
---
src/rpc/virnettlscontext.c | 149 ++++++++++++++++++++++++++++++++++++++++++-
1 files changed, 145 insertions(+), 4 deletions(-)
diff --git a/src/rpc/virnettlscontext.c b/src/rpc/virnettlscontext.c
index 4fa6fbb..b0dcfe5 100644
--- a/src/rpc/virnettlscontext.c
+++ b/src/rpc/virnettlscontext.c
@@ -67,6 +67,7 @@ struct _virNetTLSSession {
bool handshakeComplete;
+ bool isServer;
char *hostname;
gnutls_session_t session;
virNetTLSSessionWriteFunc writeFunc;
@@ -95,6 +96,134 @@ static void virNetTLSLog(int level, const char *str) {
VIR_DEBUG("%d %s", level, str);
}
+
+static gnutls_x509_crt_t virNetTLSContextSanityCheckCert(bool isServer,
+ const char *certFile)
+{
+ gnutls_datum_t data;
+ gnutls_x509_crt_t cert = NULL;
+ char *buf = NULL;
+ int ret = -1;
+ time_t now;
+
+ if ((now = time(NULL)) == ((time_t)-1)) {
+ virReportSystemError(errno, "%s",
+ _("cannot get current time"));
+ goto cleanup;
+ }
+
+ if (gnutls_x509_crt_init(&cert) < 0) {
+ virNetError(VIR_ERR_SYSTEM_ERROR, "%s",
+ _("Unable to initialize certificate"));
+ goto cleanup;
+ }
+
+ if (virFileReadAll(certFile, (1<<16), &buf) < 0)
+ goto cleanup;
+
+ data.data = (unsigned char *)buf;
+ data.size = strlen(buf);
+
+ if (gnutls_x509_crt_import(cert, &data, GNUTLS_X509_FMT_PEM) < 0) {
+ virNetError(VIR_ERR_SYSTEM_ERROR, isServer ?
+ _("Unable to import server certificate %s") :
+ _("Unable to import client certificate %s"),
+ certFile);
+ goto cleanup;
+ }
+
+ if (gnutls_x509_crt_get_expiration_time(cert) < now) {
+ virNetError(VIR_ERR_SYSTEM_ERROR, isServer ?
+ _("The server certificate %s has expired") :
+ _("The client certificate %s has expired"),
+ certFile);
+ goto cleanup;
+ }
+
+ if (gnutls_x509_crt_get_activation_time(cert) > now) {
+ virNetError(VIR_ERR_SYSTEM_ERROR, isServer ?
+ _("The server certificate %s is not yet active") :
+ _("The client certificate %s is not yet active"),
+ certFile);
+ goto cleanup;
+ }
+
+ ret = 0;
+
+cleanup:
+ if (ret != 0) {
+ gnutls_x509_crt_deinit(cert);
+ cert = NULL;
+ }
+ VIR_FREE(buf);
+ return cert;
+}
+
+
+static int virNetTLSContextSanityCheckCredentials(bool isServer,
+ const char *cacertFile,
+ const char *certFile)
+{
+ gnutls_x509_crt_t cert = NULL;
+ gnutls_x509_crt_t cacert = NULL;
+ int ret = -1;
+ unsigned int status;
+
+ if (access(certFile, R_OK) == 0) {
+ if (!(cert = virNetTLSContextSanityCheckCert(isServer, certFile)))
+ goto cleanup;
+ }
+ if (access(cacertFile, R_OK) == 0) {
+ if (!(cacert = virNetTLSContextSanityCheckCert(isServer, cacertFile)))
+ goto cleanup;
+ }
+
+ if (cert && cacert) {
+ if (gnutls_x509_crt_list_verify(&cert, 1,
+ &cacert, 1,
+ NULL, 0,
+ 0, &status) < 0) {
+ virNetError(VIR_ERR_SYSTEM_ERROR, "%s", isServer ?
+ _("Unable to verify server certificate against CA certificate") :
+ _("Unable to verify client certificate against CA certificate"));
+ goto cleanup;
+ }
+
+ if (status != 0) {
+ const char *reason = _("Invalid certificate");
+
+ if (status & GNUTLS_CERT_INVALID)
+ reason = _("The certificate is not trusted.");
+
+ if (status & GNUTLS_CERT_SIGNER_NOT_FOUND)
+ reason = _("The certificate hasn't got a known issuer.");
+
+ if (status & GNUTLS_CERT_REVOKED)
+ reason = _("The certificate has been revoked.");
+
+#ifndef GNUTLS_1_0_COMPAT
+ if (status & GNUTLS_CERT_INSECURE_ALGORITHM)
+ reason = _("The certificate uses an insecure algorithm");
+#endif
+
+ virNetError(VIR_ERR_SYSTEM_ERROR,
+ _("Our own certificate %s failed validation against %s: %s"),
+ certFile, cacertFile, reason);
+ goto cleanup;
+ }
+ }
+
+ ret = 0;
+
+cleanup:
+ if (cert)
+ gnutls_x509_crt_deinit(cert);
+ if (cacert)
+ gnutls_x509_crt_deinit(cacert);
+ return ret;
+}
+
+
static int virNetTLSContextLoadCredentials(virNetTLSContextPtr ctxt,
bool isServer,
const char *cacert,
@@ -217,6 +346,10 @@ static virNetTLSContextPtr virNetTLSContextNew(const char *cacert,
goto error;
}
+ if (requireValidCert &&
+ virNetTLSContextSanityCheckCredentials(isServer, cacert, cert) < 0)
+ goto error;
+
if (virNetTLSContextLoadCredentials(ctxt, isServer, cacert, cacrl, cert, key) < 0)
goto error;
@@ -574,15 +707,21 @@ static int virNetTLSContextValidCertificate(virNetTLSContextPtr ctxt,
}
if (gnutls_x509_crt_get_expiration_time(cert) < now) {
- virNetError(VIR_ERR_SYSTEM_ERROR, "%s",
- _("The client certificate has expired"));
+ /* Warning is reversed from what you expect, since with
+ * this code it is the Server checking the client and
+ * vica-verca */
+ virNetError(VIR_ERR_SYSTEM_ERROR, "%s", sess->isServer ?
+ _("The client certificate has expired") :
+ _("The server certificate has expired"));
gnutls_x509_crt_deinit(cert);
goto authdeny;
}
if (gnutls_x509_crt_get_activation_time(cert) > now) {
- virNetError(VIR_ERR_SYSTEM_ERROR, "%s",
- _("The client certificate is not yet active"));
+ /* client/server order reversed. see above */
+ virNetError(VIR_ERR_SYSTEM_ERROR, "%s", sess->isServer ?
+ _("The client certificate is not yet active") :
+ _("The server certificate is not yet active"));
gnutls_x509_crt_deinit(cert);
goto authdeny;
}
@@ -756,6 +895,8 @@ virNetTLSSessionPtr virNetTLSSessionNew(virNetTLSContextPtr ctxt,
gnutls_transport_set_pull_function(sess->session,
virNetTLSSessionPull);
+ sess->isServer = ctxt->isServer;
+
return sess;
error:
--
1.7.4.4
13 years, 4 months
Re: [libvirt] [virt-tools-list] [PATCH virt-manager v4] Add inspection to virt-manager
by Richard W.M. Jones
On Tue, Jul 19, 2011 at 10:46:49AM -0400, Cole Robinson wrote:
> Maybe we can cache the png data per detected OS value rather than per
> VM? Not sure if that collides with licensing issues, but would likely
> mean storing less data on disk.
You can't do that without getting into trademark issues. The icon
that is displayed must have come from precisely the same guest.
The icons are not large anyhow.
One thing I meant to ask about Dan's proposal:
> > $HOME/.local/libvirt/$CONN_URI/$DOMAIN_UUID/screenshot.png
> > $HOME/.local/libvirt/$CONN_URI/$DOMAIN_UUID/icon.png
> > $HOME/.local/libvirt/$CONN_URI/$DOMAIN_UUID/osinfo.json
Do we need the connection URI? Isn't the dom UUID unique enough?
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming blog: http://rwmj.wordpress.com
Fedora now supports 80 OCaml packages (the OPEN alternative to F#)
http://cocan.org/getting_started_with_ocaml_on_red_hat_and_fedora
13 years, 4 months
[libvirt] [PATCH] Fix reporting of errors for p2p migration
by Daniel P. Berrange
Starting/ending jobs when closing the connection may reset any
error which was reported earlier in p2p migration. We must
save the original error before doing so. This means we can also
just call virConnectClose as normal, instead of virUnrefConnect
* src/qemu/qemu_migration.c: Preserve errors in p2p migration
---
src/qemu/qemu_migration.c | 10 ++++++++--
1 files changed, 8 insertions(+), 2 deletions(-)
diff --git a/src/qemu/qemu_migration.c b/src/qemu/qemu_migration.c
index dfa80e3..1803b9f 100644
--- a/src/qemu/qemu_migration.c
+++ b/src/qemu/qemu_migration.c
@@ -2141,6 +2141,8 @@ static int doPeer2PeerMigrate(struct qemud_driver *driver,
int ret = -1;
virConnectPtr dconn = NULL;
bool p2p;
+ virErrorPtr orig_err = NULL;
+
VIR_DEBUG("driver=%p, sconn=%p, vm=%p, xmlin=%s, dconnuri=%s, "
"uri=%s, flags=%lx, dname=%s, resource=%lu",
driver, sconn, vm, NULLSTR(xmlin), NULLSTR(dconnuri),
@@ -2191,10 +2193,14 @@ static int doPeer2PeerMigrate(struct qemud_driver *driver,
dconnuri, flags, dname, resource);
cleanup:
- /* don't call virConnectClose(), because that resets any pending errors */
+ orig_err = virSaveLastError();
qemuDomainObjEnterRemoteWithDriver(driver, vm);
- virUnrefConnect(dconn);
+ virConnectClose(dconn);
qemuDomainObjExitRemoteWithDriver(driver, vm);
+ if (orig_err) {
+ virSetError(orig_err);
+ virFreeError(orig_err);
+ }
return ret;
}
--
1.7.4.4
13 years, 4 months
