February 2013 - Devel - Libvirt List Archives

[libvirt] [PATCHv3 0/5] snapshot revert-and-create

by Eric Blake

v2 was here: https://www.redhat.com/archives/libvir-list/2012-November/msg00818.html One patch from v2 has already been committed. This patch additionally adds a qemu implementation for the new flag, and I have tested creation of offline branches (I still need to test creation of a branch from a disk-only or online external snapshot). I've also started documenting my plans for a new revert FOLLOW flag, which updates to the current state of external files (rather than reverting completely to … [View More]

12 years, 1 month

2
6
0 / 0

[libvirt] [PATCH v2] qemu: Don't fail to shutdown domains with unresponsive agent

by Michal Privoznik

Currently, qemuDomainShutdownFlags() chooses the agent method of shutdown whenever the agent is configured. However, this assumption is not enough as the guest agent may be unresponsive at the moment. So unless guest agent method has been explicitly requested, we should fall back to the ACPI method. --- diff to v1: - Rework some conditions as Eric suggested in v1 src/qemu/qemu_driver.c | 38 ++++++++++++++++++++++---------------- 1 file changed, 22 insertions(+), 16 deletions(-) diff --git … [View More]a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 1e96915..1e96aa4 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -1702,40 +1702,40 @@ static int qemuDomainShutdownFlags(virDomainPtr dom, unsigned int flags) { virDomainObjPtr vm; int ret = -1; qemuDomainObjPrivatePtr priv; - bool useAgent = false; + bool useAgent = false, agentRequested, acpiRequested; virCheckFlags(VIR_DOMAIN_SHUTDOWN_ACPI_POWER_BTN | VIR_DOMAIN_SHUTDOWN_GUEST_AGENT, -1); - /* At most one of these two flags should be set. */ - if ((flags & VIR_DOMAIN_SHUTDOWN_ACPI_POWER_BTN) && - (flags & VIR_DOMAIN_SHUTDOWN_GUEST_AGENT)) { - virReportInvalidArg(flags, "%s", - _("flags for acpi power button and guest agent are mutually exclusive")); - return -1; - } - if (!(vm = qemuDomObjFromDomain(dom))) goto cleanup; priv = vm->privateData; + agentRequested = flags & VIR_DOMAIN_SHUTDOWN_GUEST_AGENT; + acpiRequested = flags & VIR_DOMAIN_SHUTDOWN_ACPI_POWER_BTN; - if ((flags & VIR_DOMAIN_SHUTDOWN_GUEST_AGENT) || - (!(flags & VIR_DOMAIN_SHUTDOWN_ACPI_POWER_BTN) && - priv->agent)) + /* Prefer agent unless we were requested to not to. */ + if (agentRequested || (!flags && priv->agent)) useAgent = true; - if (useAgent) { - if (priv->agentError) { + if (priv->agentError) { + if (agentRequested && !acpiRequested) { virReportError(VIR_ERR_AGENT_UNRESPONSIVE, "%s", _("QEMU guest agent is not " "available due to an error")); goto cleanup; + } else { + useAgent = false; } - if (!priv->agent) { + } + + if (!priv->agent) { + if (agentRequested && !acpiRequested) { virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s", _("QEMU guest agent is not configured")); goto cleanup; + } else { + useAgent = false; } } @@ -1752,7 +1752,13 @@ static int qemuDomainShutdownFlags(virDomainPtr dom, unsigned int flags) { qemuDomainObjEnterAgent(vm); ret = qemuAgentShutdown(priv->agent, QEMU_AGENT_SHUTDOWN_POWERDOWN); qemuDomainObjExitAgent(vm); - } else { + } + + /* If we are not enforced to use just an agent, try ACPI + * shutdown as well in case agent did not succeed. + */ + if (!useAgent || + (ret < 0 && (acpiRequested || !flags))) { qemuDomainSetFakeReboot(driver, vm, false); qemuDomainObjEnterMonitor(driver, vm); -- 1.8.1.4 [View Less]

12 years, 1 month

2
4
0 / 0

[libvirt] a script for libvirt-llxc + systemd on Fedora18

by Kamezawa Hiroyuki

Hi. At playing libvirt-lxc on Fedora18, I found that the user needs some workarounds to run /sbin/init as the root process of container. With Fedora15, I found this https://gist.github.com/peo3/1142202. And I know virt-sandbox-service have some clever way to understand RPM to make domain running with a service. Here is a script I wrote for F18 inspired by above two. Creating LXC domain running systemd as /init program. I'd like to get some feedback....and ...can't this kind of scripts be … [View More]maintained by some virt-tools project ? Anyway, playing with lxc is fun. I'm glad if someone makes this script better :) == A user can create a lxc container by # ./lxc_systemd_setup.py -r -n Test This will generate my.xml and lxc file tree on /opt/lxc/Test # virsh -c lxc:/// define my.xml # virsh -c lxc:/// start my.xml This domain has following characteristics - /usr is bind mounted to host's /usr - all files other than files in /usr are copied - systemd and pam modules are tweaked a bit - eth0 is up by dhcp. - systemd, rsyslog, sshd is running. Users can add rpm package as # ./lxc_systemd_setup.py -r -n Test -a httpd some options may not work because of my bad skills. ==lxc_systemd_setup.py== #!/usr/bin/python # # A scripts for creating LXC domain where some daemons are running under systemd. # Tested with Fedora18, running systemd, rsyslog, sshd in a container at default. # # Most config files, including all security settings as passwd, selinux, # ssl/pki files are copied from the host. So, please fix them before # running a domain. # # New root dir will be /opt/lxc/<domain name> at default and domain XML # definition will be saved as my.xml # # All instllation under /usr is shared among containres/host, binaries are # not copied at all. /etc and /var are copied. So, you can adjust config files # under container as you like. # # for easy creation, run # # lxc_systemd_setup.py -n Test -r # # virsh -c lxc:/// define my.xml # # virsh -c lxc:/// start Test --console # # This will build a lxc root filesystem under /opt/lxc/Test and copy # required files under it. /usr will be shared with the host. # # to add some packages, pass package name with '-a' option. But this script doesn't # handle dependency of RPM, at all. please take care of it. # # for running httpd. # # lxc_systemd_setup.py -n Test -r -a httpd # _and_ you need to fix hostname lookup problem to run httpd. i.e. you may need to fix # some files under /etc.... # # please see # # lxc_systemd_setup.py --help # for other options. some may not work ;) # import sys, os, shutil from subprocess import Popen, PIPE from optparse import OptionParser,OptionGroup import re import rpm parser = OptionParser() parser.add_option('-p', '--prefix', action='store', default='/opt/lxc', type='string', dest='pathname', help='prefix of guest root path') parser.add_option('-n', '--name', action='store', default='_unknown', type='string', dest='domain_name', help='name of domain') parser.add_option('-o', '--out', action='store', default='my.xml', type='string', dest='def_file', help='name of generated xml def') parser.add_option('-r', '--renew', action='store_true', default=False, dest='renew_tree', help='delete existing tree if exists') parser.add_option('-D', '--destroy', action='store_true', default=False, dest='destroy_tree', help='destroy existing tree and quit') parser.add_option('-f', '--force', action='store_true', default=False, dest='force', help='update all files without checking timestamp') parser.add_option('-a','--add-packages', action='append', type='string', dest='package_list', help='copy package config to container') parser.add_option('-s','--skip-packages', action='append', type='string', dest='skip_list', help='skip packages') parser.add_option('-m','--memory', action='store', default='1024000', type='str', dest='memory', help='memory size of domain') parser.add_option('-H','--hostname', action='store', default='lxc', type='str', dest='hostname', help='hostname of domain') (options, args) = parser.parse_args(sys.argv) # # Utility functions. # # # Remove all files under domain ROOT. # def destroy_all(path) : if (not os.path.exists(path)) : return shutil.rmtree(path, ignore_errors=True) # # Check a file in a container is newer than hosts. # def file_is_newer(a, b) : time_a = os.stat(a).st_mtime time_b = os.stat(b).st_mtime return a > b # # Check Host's distro. # def check_version() : useRPM=True version="unknown" if (os.path.exists('/etc/redhat-release')) : with open('/etc/redhat-release') as f: version_string = f.readline() if (re.match("Fedora release 18.+$", version_string)) : version = "Fedora18" else : useRPM=False return (version, useRPM) # # directories created at domain creation (tested with Fedora 18) # ROOTDIR= options.pathname + "/" + options.domain_name class InstallInfo : def __init__(self) : self.DIRS = [] self.BINDDIRS = [] self.SYMLINKS=[] self.PACKAGES = [] self.FILES = [] self.MERGED = [] def add_dirs(self, x) : if (isinstance(x, str)) : x = [x] self.DIRS = self.DIRS + x def dirs(self) : return self.DIRS def add_files(self, x) : if (isinstance(x, str)) : x = [x] self.FILES += x def files(self) : return self.FILES def add_binds(self, x) : if (isinstance(x, str)) : x = [x] self.BINDDIRS = self.BINDDIRS + x def binds(self) : return self.BINDDIRS def add_links(self, x) : self.SYMLINKS = self.SYMLINKS + x def links(self) : return self.SYMLINKS def add_packages(self, x) : self.PACKAGES = self.PACKAGES + x def packages(self) : return self.PACKAGES def merge(self) : self.DIRS = list(set(self.DIRS)) self.FILES = list(set(self.FILES)) self.BINDDIRS = sorted(list(set(self.BINDDIRS))) ret = True ents =[] paths = [] for ent in self.DIRS : ents.append((ent, 'dir', '')) paths.append(ent) for ent in self.FILES : ents.append((ent, 'file', '')) paths.append(ent) for ent in self.SYMLINKS : ents.append((ent[0], 'link', ent[1])) paths.append(ent[0]) if (len(paths) - len(list(set(paths)))) : ret = False self.MERGED = sorted(ents) return ret def merged(self) : return self.MERGED # # Gather RPM information and copy config files to proper place. # class CopyRPMHandler: def __init__(self, name) : self.name = name self.files = [] self.service = "" def verify(self) : ts = rpm.TransactionSet() mi = ts.dbMatch('name', self.name) if not mi : return False # get list of files. for h in mi : myhead = h break fi = myhead.fiFromHeader() for x in fi : self.files.append(x[0]); self.test_service() self.strip_binds() return True def paths(self) : return self.files # # check all files in RPM which are not under bind-mount. # def strip_binds(self) : # remove all ents under /usr for avoiding copy. temp = self.files self.files = [] for file in temp : if (file == '') : continue if (re.match("/usr/.+$", file)) : continue if (re.match("/bin/.+$", file)) : continue if (re.match("/lib/.+$", file)) : continue if (re.match("/sbin/.+$", file)) : continue if (re.match("/lib64/.+$", file)) : continue self.files.append(file) self.files.sort() return def test_service(self) : for file in self.files : if (re.match("/usr/lib/systemd/system.+\.service$", file)) : ent.service = file # # Functions for workarounds. # # # systemd: create our own basic.target for avoiding some startups. # def systemd_tune() : # # we need to avoid some special services by systemd. # modify basic.target and avoid them. # filename = ROOTDIR + "/etc/systemd/system/basic.target" data="""[Unit] Description=Basic System Documentation=man:systemd.special(7) Requires=systemd-tmpfiles-setup.service sockets.target After=systemd-tmpfiles-setup.service sockets.target RefuseManualStart=yes """ with open(filename,"w") as f: f.write(data) # # we need getty only with tty1 # os.symlink("/usr/lib/systemd/system/getty@.service", ROOTDIR + "/etc/systemd/system/getty.target.wants/getty(a)tty1.service") # # Create ifcfg-eth0 and add service to bring up it. # def eth0_service() : # # /etc/sysconfig/network is generated by annaconda and we cannot # find it by rpms. # filename = ROOTDIR + "/etc/sysconfig/network" shutil.copy("/etc/sysconfig/network", filename); # # ifconfig setting for eth0 # filename = ROOTDIR + "/etc/sysconfig/network-scripts/ifcfg-eth0" data="""DEVICE=eth0 BOOTPROTO=dhcp ONBOOT=yes NAME=eth0 TYPE=Ethernet """ with open(filename, "w") as f: f.write(data) print "Creating %s" % filename filename = ROOTDIR + "/etc/systemd/system/lxc-eth0.service" data="""[Unit] Before=multi-user.target Conflicts=shutdown.target Description=bring up eth0 in this container [Service] ExecStart=/usr/sbin/ifup eth0 Type=simple """ with open(filename, "w") as f: f.write(data) print "Creating %s" % filename # # Bring up this. # filename = ROOTDIR + "/etc/systemd/system/basic.target.wants/lxc-eth0.service" src = "/etc/systemd/system/lxc-eth0.service" os.symlink(src, filename) # # Make fstab empty # def empty_fstab() : filename = ROOTDIR + "/etc/fstab" with open(filename, "w") as f: f.truncate(0) # # in Fedora18, pam's pam_loginuid.so doesn't work under container # we need to disable it. # def pam_tune() : pamdir = ROOTDIR + "/etc/pam.d" for root, dirs, files in os.walk(pamdir) : for path in files : path = root + "/" + path if (os.path.islink(path)) : continue data ="" with open(path) as f: for line in f : if (re.match("^.+pam_loginuid.so.*$", line)) : line = "#" + line data += line with open(path, "w") as f: f.write(data) # # securetty set up for login via system console. # def securetty_tune() : path = ROOTDIR + "/etc/securetty" with open(path, "a") as f: f.write("pts/0\n") # # set hostname of guest domain. # def hostname_modify() : path = ROOTDIR + "/etc/hostname" with open(path, "w") as f: f.write(options.hostname + "\n") # # parse memory size. # def parse_memory(data) : if data[-1] == 'K' : x = int(data[0:-1]) return str(x * 1024) elif data[-1] == 'M' : x = int(data[0:-1]) return str(x * 1024 * 1024) elif data[-1] == 'G' : x = int(data[0:-1]) return str(x * 1024 * 1024 * 1024) else : return data # # Main routine starts here ! # version, useRPM = check_version() if (not useRPM) : print 'now, we can handle RPM only' exit(1) info = InstallInfo() # Build a information. # # At first, gather required RPM information and some tweaks for distro. # if (version == 'Fedora18') : # # now, dont'handle yum and rpm info in container, so create fake dirs # instead of copying yum info by RPM. # info.add_dirs(["/etc/yum", "/etc/yum/protected.d", "/etc/yum/pluginconf.d","/etc/yum/vars"]) # # We share /usr between host and guest. # info.add_binds(["/usr"]) # # For Fedora18, we need following copies of configs packages at least. # info.add_packages(["filesystem","setup","rpm", "selinux-policy"]) info.add_packages(["systemd", "dbus", "initscripts","util-linux"]) info.add_packages(["pam","passwd", "crontabs","kmod","logrotate","rsyslog"]) info.add_packages(["openssh","openssh-server", "chkconfig","authconfig"]) info.add_packages(["glibc", "mailcap"]) # symlink and dirs for systemd info.add_links([["/etc/systemd/system/default.target", "/lib/systemd/system/multi-user.target"]]) info.add_dirs(["/etc/systemd/system/basic.target.wants"]) info.add_dirs(["/etc/systemd/system/default.target.wants"]) info.add_dirs(["/etc/systemd/system/getty.target.wants"]) info.add_dirs(["/etc/systemd/system/multi-user.target.wants"]) info.add_dirs(["/etc/systemd/system/sockets.target.wants"]) info.add_dirs(["/etc/systemd/system/sysinit.target.wants"]) info.add_dirs(["/etc/systemd/system/system-update.target.wants"]) # # Merge package list # package_names = info.packages() if (options.package_list) : package_names += options.package_list # Uniq. package_names = list(set(package_names)) # # delete unnecessary packages from list. # if (options.skip_list) : for name in options.skip_list : if (name in package_names) : package_names.remove(name) # # Verify package list (check installation of packages) # packages = [] error = False if (useRPM) : for name in package_names : ent = CopyRPMHandler(name) if (ent.verify()) : packages.append(ent) else : print "Couldn't find a package [%s] in RPM DB." % (name) error = True if (error) : exit(1) # # Now, we confirmed all RPMS required are installed in the host. # service_files = [] # # Extract dir,symlink,file information from RPMS. Later, we'll copy all # files other than /usr. # for ent in packages : for path in ent.paths() : if (not os.path.exists(path)) : continue if (os.path.islink(path)) : src = os.readlink(path) info.add_links([[path, src]]) elif (os.path.isfile(path)) : info.add_files(path) elif (os.path.isdir(path)) : info.add_dirs(path) if (ent.service != ""): service_files.append(ent.service) # # Uniq and sort it. # if (not info.merge()) : print "some confilction of files may happen..." # # Check Domain name is passed. # if (options.domain_name == '_unknown') : print "Guest Domain name must be specified" exit(1) # # Destroy tree. # if (options.destroy_tree) : destroy_all(ROOTDIR) exit(0) # # At first, clear tree if required. # (*) the scirpt may not work if we don't destroy the tree .... # if (os.path.exists(ROOTDIR)) : if (options.renew_tree) : destroy_all(ROOTDIR) # Create root dir try: os.mkdir(ROOTDIR) except: print "cannot create root dir %s" % ROOTDIR exit(1) # Ok, make world based on information gathered from RPMS. for ents in info.merged() : guestpath = ROOTDIR + ents[0] try: if (ents[1] == 'dir') : if (not os.path.exists(guestpath)) : print "Creating dir %s" % (guestpath) os.makedirs(guestpath) elif (ents[1] == 'link') : if (not os.path.exists(guestpath)) : print "Creating symlink %s => %s" % (guestpath, ents[2]) os.symlink(ents[2], guestpath) elif (ents[1] == 'file') : if (options.force or not os.path.exists(guestpath) or file_is_newer(ents[0], guestpath)) : print "Copyfile %s" % (guestpath) shutil.copy(ents[0], guestpath) except: print "error at creating tree %s" % guestpath exit(1) # # setup service files if necessary. # for file in service_files : service = os.path.basename(file) p = re.compile('WantedBy=(.+)$') for line in open(file, 'r') : m = p.match(line) if (m) : target = m.group(1) pathname = ROOTDIR + "/etc/systemd/system/" + target + ".wants/" + service print "%s=>%s" % (pathname, file) os.symlink(file, pathname) # # Tweak system settings. # if (version == "Fedora18") : # diable some services. dir = ROOTDIR + "/etc/systemd/system/" os.symlink("/dev/null", dir + "sysinit.target") os.symlink("/dev/null", dir + "console-shell.service") os.symlink("/dev/null", dir + "fedora-readonly.service") os.symlink("/dev/null", dir + "fedora-storage-init.service") systemd_tune() # modify basic.target etc... eth0_service() # bringup eth0 without udev empty_fstab() # make /etc/fstab empty pam_tune() # disable some pam module securetty_tune() # add pts/0 to securetty hostname_modify() # # Generate a Domain Def. # domain = r""" <domain type='lxc'> <name>%(NAME)s</name> <memory unit='bytes'>%(MEMORY)s</memory> <vcpu>1</vcpu> <os> <type arch='x86_64'>exe</type> <init>/sbin/init</init> </os> <clock offset='utc'/> <on_poweroff>destroy</on_poweroff> <on_reboot>restart</on_reboot> <on_crash>destroy</on_crash> <devices> <emulator>/usr/libexec/libvirt_lxc</emulator> <filesystem type='mount' accessmode='passthrough'> <source dir='%(ROOTDIR)s'/> <target dir='/'/> </filesystem> <filesystem type='mount' accessmode='passthrough'> <source dir='/usr'/> <target dir='/usr'/> </filesystem> <filesystem type='ram'> <source usage='%(MEMORY)s'/> <target dir='/tmp'/> </filesystem> <filesystem type='ram'> <source usage='%(MEMORY)s'/> <target dir='/dev/shm'/> </filesystem> <interface type="network"> <source network="default"/> </interface> <console type='pty'> <target type='lxc' port='0'/> </console> </devices> </domain> """ % {'NAME':options.domain_name, 'MEMORY': parse_memory(options.memory), 'ROOTDIR':ROOTDIR} with open(options.def_file, "w") as f : f.write(domain) [View Less]

12 years, 1 month

3
2
0 / 0

[libvirt] [PATCH] qemu: virConnectGetVersion returns bogus value

by Viktor Mihajlovski

The unitialized local variable qemuVersion can cause an random value to be returned for the hypervisor version, observable with virsh version. Introduced by commit b46f7f4a0b96c2d2d01d64d960bd7bc90dc16b0c Signed-off-by: Viktor Mihajlovski <mihajlov(a)linux.vnet.ibm.com> --- src/qemu/qemu_driver.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 0f6a431..825babd 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/… [View More]

12 years, 1 month

2
1
0 / 0

[libvirt] [PATCH 1/1] Remove contiguous CPU indexes assumption

by Li Zhang

From: Li Zhang <zhlcindy(a)linux.vnet.ibm.com> When getting CPUs' information, it assumes that CPU indexes are not contiguous. But for ppc64 platform, CPU indexes are not contiguous because SMT is needed to be disabled, so CPU information is not right on ppc64 and vpuinfo, vcpupin can't work corretly. This patch is to remove the assumption to be compatible with ppc64. Test: 4 vcpus are assigned to one VM and execute vcpuinfo command. Without patch: There is only one vcpu … [View More]

12 years, 1 month

2
4
0 / 0

[libvirt] [PATCH 0/4] Add cpuset support for LXC

by Gao feng

This patchset intend to add cpuset support for LXC. in order to don't create too many redundant codes, this patchset also rename some functions. Gao feng (4): rename qemuGetNumadAdvice to virDomainGetNumadAdvice LXC: allow uses advisory nodeset from querying numad remove the redundant codes LXC: add cpuset support for lxc src/conf/domain_conf.c | 144 ++++++++++++++++++++++++++++++++++++++++++ src/conf/domain_conf.h | 4 ++ src/libvirt_private.syms | 3 +- src/lxc/lxc_cgroup.… [View More]

12 years, 1 month

2
9
0 / 0

[libvirt] [PATCH] maint: fix typo in network docs

by Eric Blake

* docs/formatnetwork.html.in: Spell variation correctly. --- Pushing under the trivial rule. docs/formatnetwork.html.in | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/formatnetwork.html.in b/docs/formatnetwork.html.in index 41a83fa..4dd0415 100644 --- a/docs/formatnetwork.html.in +++ b/docs/formatnetwork.html.in @@ -780,7 +780,7 @@ </network></pre> <p> - Below is another IPv6 varition. Instead of a dhcp range being + … [View More]

12 years, 1 month

1
0
0 / 0

[libvirt] Compile libvirt V1.0.1 error

by harryxiyou

Hi all, When i compiled libvirt like following 1. git clone git://libvirt.org/libvirt.git；cd libvirt 2. git reset --hard v1.0.1 3. wget http://cloudxy.googlecode.com/svn/branches/hlfs/person/harry/hlfs/patches... 4. wget http://cloudxy.googlecode.com/svn/branches/hlfs/person/harry/hlfs/patches... 5. git apply hlfs_driver_for_libvirt_network_disk.patch 6. git apply hlfs_driver_for_libvirt_add_classpath.patch 7. ./autogen.sh 8. ./configure 9. make i got following errors. [..… [View More]

12 years, 1 month

2
4
0 / 0

Re: [libvirt] [Dnsmasq-discuss] Chaining instances?

by Laine Stump

This discussion should really be taking place on libvir-list - I'm Cc'ing it there. On 02/24/2013 04:11 PM, TJ wrote: > On 24/02/13 19:19, Laine Stump wrote: >> On 02/24/2013 05:09 AM, TJ wrote: >>> I wondered if maybe configuring the libvirtd dnsmasq instances to be dhcp proxies for the LAN dnsmasq, and use multiple dhcp-range's with tags might do it? >>> >>> My brain is a bit fried right now having had to fix several bugs in vmbuilder and libvirt, plus … [View More]add new functionality to libvirtd to allow its dnsmasq instance to read a conf-file and use a separate log-facility, >> What you're talking about doing sounds *very* useful to have supported >> directly in libvirt, but you may want to contact libvirt's upstream >> developers (at libvir-list(a)redhat.com, or on irc.oftc.net in #virt) >> prior to expending a lot of effort on libvirt code - in general a >> problem may already be solved in a different manner, or somebody else >> may already be working on it. Failing that, there may have already been >> considerable debate on a particular subject, and a path to a solution >> generally agreed on, but with nobody yet working on the code. > It turns out that what I need(ed) to do was completely *disable* libvirt's use of dnsmasq > and instead use Simon's related "dhcp-helper" program which acts as a full dhcp-relay using, f.e: > > /usr/sbin/dhcp-helper -u libvirt-dnsmasq -i virbr1 -b eth0 > > This forwards all requests to the primary dnsmasq DHCP server on the LAN which matches against the > appropriate dhcp-range: > > # physical network leases > dhcp-range=set:phys,10.254.251.50,10.254.251.199,255.255.255.0,1440m > # subnet for VMs on server1 > dhcp-range=set:vmlan1,10.254.1.100,10.254.1.199,255.255.255.0,1440m > # default route (gateway) > dhcp-option=tag:phys,option:router,10.254.251.1 > dhcp-option=tag:vmlan1,option:router,10.254.1.1 > # DNS server > dhcp-option=6,10.254.251.1 > > > To that end this evening I added two additional options to libvirt: > > <dnsmasq enabled='true'/> > <dhcphelper enabled='false'/> libvirt's XML should remain implementation-agnostic. We don't want to have an option to disable the specific implementation of dns+dhcp called "dnsmasq"; rather we want to be able to enable/disable a network's dhcp server and/or dns server. (The reason for this is that we want to leave the door open for someone to implement a different backend using a different dhcp server and/or dns server and be able to use the same configuration.) We have actually previously discussed disabling dns and/or dhcp (see http://www.redhat.com/archives/libvir-list/2012-November/msg00861.html). It is already possible to completely disable dhcp - simply don't include a <dhcp> element in the network definition. As for dns, from the very beginning dns has been *always* enabled on all networks, so even though there is a <dns> element, simply having not having one is not a valid way to say "no dns server" (as it would cause backward compatibility problems with existing installations). Instead, the conclusion of the above-mentioned thread was that the proper way to handle this would be to add an "enable" element to the existing <dns> element, which would default to "yes". To disable libvirt-supplied dns services for a subnet, you would just put: <dns enable='no'/> in the network definition. In the case that dns had enable='no' AND there was no <dhcp> element, dnsmasq simply wouldn't be started at all. That hasn't been implemented yet, but shouldn't be too complex to do. As for dhcp-helper, aside from the fact that again you've created an option that is specific to a particular implementation of what is needed, that command isn't universally available anyway (it's not in any package for Fedora/RHEL, for example), and I'm not sure that it's really necessary to have it started up by libvirt anyway - can it detect newly created/upped interfaces as dnsmasq can? If so, it could be started up by regular host system config even before libvirt was started. > > These are the default values when the options are *not* defined. They allow the admin to disable dnsmasq entirely: > > <dnsmasq enabled='false'/> > > and to enable dhcp-helper: > > <dhcphelper enabled='true'/> > > Using two new functions: > > int networkBuildDhcphelperArgv(...) > int networkBuildDhcpHelperCommandLine(...) > > the existing function: > > int networkStartDhcpDaemon(...) > > is able to launch either or both of dnsmasq and dhcp-helper with the correct options. > > dhcp-helper's "-i" option is filled from network->def->bridge and its "-b" option is taken from > the first forward device declared in a <forward dev='?'/> or <interface dev='?'/>. If no forward > device is declared it throws a VIR_ERR_INVALID_INTERFACE error with appropriate explanatory text. > > > [View Less]

12 years, 1 month

2
4
0 / 0

[libvirt] q35 machine type and libvirt.

by Laine Stump

Now that qemu is getting the q35 machine type, libvirt needs to support it. As far as I understand, from libvirt's point of view, q35 is just another x86_64 system, but with a different set of implicit devices, and possibly some extra rules limiting which devices can be plugged into which bus/slot on the guest. That means that in order to support it, we will need to recognize when a q35-based machine type is being created, and auto-add all the implicit devices to libvirt's config model for … [View More]that domain, then pay attention to the extra rules when assigning addresses for all the user-added devices. We already add implicit controllers/devices for pc-based machine types; as a matter of fact, currently, libvirt improperly assumes (for the purposes of adding implicit devices) that *every* virtual machine is based on the "pc" machine type (or rather it just doesn't pay attention), so it always adds all the implicit devices for a pc machine type for every domain. This of course is already incorrect for many (probably all?) non-x86 machine types, even before we add q35 into the mix. To fix this, it might be reasonable (and arguably, it's necessary to fix the problem in a backward-compatible manner) to just setup a table of machinetype ==> implicit device lists, look up the machine type in this table, and add the devices needed for that machine type. This goes against libvirt's longstanding view of machinetype as being an opaque value that it merely passes through to qemu, but it's manageable for the existing machine types (even including q35), since it's a finite set. But it starts to be a pain to maintain when you think about future additions - yet another case where new functionality in qemu will require an update to libvirt before it can be fully used via libvirt. In the long term, it would be very useful / more easily maintainable to have a qemu status command available via QMP which would return the list of implicit devices (and their PCI addresses) for any requested machine type. It would be necessary that this command be callable multiple times within a single execution of qemu, giving it a different machinetype each time. This way libvirt could first query the list of available machinetypes in this particular qemu binary, then request the list of implicit devices for each machine type (libvirt runs each available qemu binary *once* the first time it's requested, and caches all such capabilities information so that it doesn't need to re-run qemu again and again). My limited understanding of qemu's code is that qemu itself doesn't have a table of this information as data, but instead has lines of code that are executed to create it, thus making it impractical to provide the list of devices for a machinetype without actually instantiating a machine of that type. What's the feasibility of adding such a capability (and in the process likely making the list of implicit devices in qemu itself table/data driven rather than constructed with lines of code). More questions: 1) It seems that the exact list of devices for the basic q35 machine type hasn't been settled on yet, is that correct? 2) Are there other issues aside from implicit controller devices I need to consider for q35? For example, are there any devices that (as I recall is the case for some devices on "pc") may or may not be present, but if they are present they are always at a particular PCI address (meaning that address must be reserved)? I've also just learned that certain types of PCIe devices must be plugged into certain locations on the guest bus? ("root complex" devices - is there a good source of background info to learn the meaning of terms like that, and the rules of engagement? libvirt will need to know/follow these rules.) 3) What new types of devices/controllers must be supported for a properly functioning q35 machine? [View Less]

12 years, 1 month

6
8
0 / 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

Devel February 2013