- Devel - libvirt List Archives

[libvirt] Clock problems on live migration
by Paul Boven 02 Apr '14

02 Apr '14

Hi everyone, When doing a live migration, Linux guests will frequently get stuck and become unresponsive, while the CPU utilization on the host for that guest goes to 100%. If you wait long enough, the guest seems to always recover, with a dmesg entry such as: Clocksource tsc unstable (delta = 35882846234 ns) So the TSC did a jump of nearly 36 seconds. I've actually had a (test-vm) stay stuck for over 11 minutes before becoming responsive again (delta = 662463064082 ns). Migrations often fail when going from server A to B, but will then work fine in the other direction. Both servers and guests are locked to the same NTP servers, and are well within 1ms from one another. Both hosts are running Ubuntu 13.04 with these versions (from Ubuntu packages): Kernel: 3.8.0-35-generic x86_64 Libvirt: 1.0.2 Qemu: 1.4.0 Gluster-fs: 3.4.2 (libvirt access the images via the filesystem, not using libgfapi yet). The interconnect between both machines (both for migration and gluster) is 10GbE. We have different guests (all Ubuntu releases, 13.04 and 13.10), and they all seem to be affected. Clocksource: kvm-clock on all guests. Clock entry from the guest XML: <clock offset='utc'/> Now as far as I've udnerstood the documentation of kvm-clock, it specifically supports live migrations, so I'm a bit surprised by these problems. There isn't all that much information to find on this issue, although I have found postings by others that seem to have run into the same issues, but without a solution. Any help would be much appreciated. Regards, Paul Boven. -- Paul Boven <boven(a)jive.nl> +31 (0)521-596547 Unix/Linux/Networking specialist Joint Institute for VLBI in Europe - www.jive.nl VLBI - It's a fringe science

1 0

[libvirt] [PATCHv2] Fix coverity-reported leak in virSecurityManagerGenLabel
by Ján Tomko 02 Apr '14

02 Apr '14

Coverity complains about a possible leak of seclabel if !sec_managers[i]->drv->domainGenSecurityLabel is true and the seclabel might be overwritten by the next iteration of the loop. This leak should never happen, because every security driver has domainGenSecurityLabel defined. --- v1: 'Remove useless NULL check in virSecurityManagerGenLabel' https://www.redhat.com/archives/libvir-list/2014-April/msg00087.html src/security/security_manager.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/src/security/security_manager.c b/src/security/security_manager.c index d68c7e9..79edb07 100644 --- a/src/security/security_manager.c +++ b/src/security/security_manager.c @@ -514,6 +514,8 @@ int virSecurityManagerGenLabel(virSecurityManagerPtr mgr, if (!sec_managers[i]->drv->domainGenSecurityLabel) { virReportUnsupportedError(); + virSecurityLabelDefFree(seclabel); + seclabel = NULL; } else { /* The seclabel must be added to @vm prior calling domainGenSecurityLabel * which may require seclabel to be presented already */ -- 1.8.3.2

2 2

[libvirt] [PATCH v3] Introduce --without-pm-utils to get rid of pm-is-supported dependency
by Cédric Bosdonnat 02 Apr '14

02 Apr '14

This uses the dbus api of systemd to check the power management capabilities of the node. --- Diff with v2: * Fixed a few dbus call problems configure.ac | 11 +++++++++ libvirt.spec.in | 9 ++++++++ src/libvirt_private.syms | 3 +++ src/util/virnodesuspend.c | 32 +++++++++++++++++++++++++ src/util/virsystemd.c | 59 +++++++++++++++++++++++++++++++++++++++++++++++ src/util/virsystemd.h | 6 +++++ 6 files changed, 120 insertions(+) diff --git a/configure.ac b/configure.ac index 73efffa..807cf0f 100644 --- a/configure.ac +++ b/configure.ac @@ -563,6 +563,10 @@ AC_ARG_WITH([chrdev-lock-files], [location for UUCP style lock files for character devices (use auto for default paths on some platforms) @<:@default=auto@:>@])]) m4_divert_text([DEFAULTS], [with_chrdev_lock_files=auto]) +AC_ARG_WITH([pm-utils], + [AS_HELP_STRING([--with-pm-utils], + [use pm-utils for power management @<:@default=yes@:>@])]) +m4_divert_text([DEFAULTS], [with_pm_utils=yes]) dnl dnl in case someone want to build static binaries @@ -1621,6 +1625,12 @@ fi AM_CONDITIONAL([WITH_PHYP],[test "$with_phyp" = "yes"]) +dnl Should we build with pm-utils support? +if test "$with_pm_utils" = "yes"; then + AC_DEFINE_UNQUOTED([WITH_PM_UTILS], 1, [whether to use pm-utils]) +fi +AM_CONDITIONAL([WITH_PM_UTILS], [test "$with_pm_utils" = "yes"]) + dnl virsh libraries VIRSH_LIBS="$VIRSH_LIBS $READLINE_LIBS" AC_SUBST([VIRSH_LIBS]) @@ -2845,6 +2855,7 @@ AC_MSG_NOTICE([ rbd: $LIBRBD_LIBS]) else AC_MSG_NOTICE([ rbd: no]) fi +AC_MSG_NOTICE([pm-utils: $with_pm_utils]) AC_MSG_NOTICE([]) AC_MSG_NOTICE([Test suite]) diff --git a/libvirt.spec.in b/libvirt.spec.in index eab9b23..5c20955 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -132,6 +132,7 @@ %define with_libssh2 0%{!?_without_libssh2:0} %define with_wireshark 0%{!?_without_wireshark:0} %define with_systemd_daemon 0%{!?_without_systemd_daemon:0} +%define with_pm_utils 1 # Non-server/HV driver defaults which are always enabled %define with_sasl 0%{!?_without_sasl:1} @@ -182,6 +183,7 @@ %if 0%{?fedora} >= 17 || 0%{?rhel} >= 7 %define with_systemd 1 %define with_systemd_daemon 1 + %define with_pm_utils 0 %endif # Fedora 18 / RHEL-7 are first where firewalld support is enabled @@ -1138,8 +1140,10 @@ Requires: nc Requires: gettext # Needed by virt-pki-validate script. Requires: gnutls-utils +%if %{with_pm_utils} # Needed for probing the power management features of the host. Requires: pm-utils +%{endif} %if %{with_sasl} Requires: cyrus-sasl # Not technically required, but makes 'out-of-box' config @@ -1395,6 +1399,10 @@ driver %define _without_systemd_daemon --without-systemd-daemon %endif +%if ! %{with_pm_utils} + %define _without_pm_utils --without-pm-utils +%endif + %define when %(date +"%%F-%%T") %define where %(hostname) %define who %{?packager}%{!?packager:Unknown} @@ -1471,6 +1479,7 @@ rm -f po/stamp-po %{?_with_firewalld} \ %{?_without_wireshark} \ %{?_without_systemd_daemon} \ + %{?_without_pm_utils} \ %{with_packager} \ %{with_packager_version} \ --with-qemu-user=%{qemu_user} \ diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index 38fbf63..ce51bdf 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -1879,6 +1879,9 @@ virSysinfoSetup; # util/virsystemd.h +virSystemdCanHibernate; +virSystemdCanHybridSleep; +virSystemdCanSuspend; virSystemdCreateMachine; virSystemdMakeMachineName; virSystemdMakeScopeName; diff --git a/src/util/virnodesuspend.c b/src/util/virnodesuspend.c index 8088931..ba4a338 100644 --- a/src/util/virnodesuspend.c +++ b/src/util/virnodesuspend.c @@ -22,6 +22,9 @@ #include <config.h> #include "virnodesuspend.h" +#ifndef WITH_PM_UTILS +# include "virsystemd.h" +#endif #include "vircommand.h" #include "virthread.h" #include "datatypes.h" @@ -260,6 +263,7 @@ int nodeSuspendForDuration(unsigned int target, * * Returns 0 if the query was successful, -1 on failure. */ +#ifdef WITH_PM_UTILS static int virNodeSuspendSupportsTarget(unsigned int target, bool *supported) { @@ -300,6 +304,34 @@ virNodeSuspendSupportsTarget(unsigned int target, bool *supported) virCommandFree(cmd); return ret; } +#else /* ! WITH_PM_UTILS */ +static int +virNodeSuspendSupportsTarget(unsigned int target, bool *supported) +{ + int ret = -1; + + if (virNodeSuspendInitialize() < 0) + return -1; + + *supported = false; + + switch (target) { + case VIR_NODE_SUSPEND_TARGET_MEM: + ret = virSystemdCanSuspend(supported); + break; + case VIR_NODE_SUSPEND_TARGET_DISK: + ret = virSystemdCanHibernate(supported); + break; + case VIR_NODE_SUSPEND_TARGET_HYBRID: + ret = virSystemdCanHybridSleep(supported); + break; + default: + return ret; + } + + return ret; +} +#endif /* WITH_PM_UTILS */ /** * virNodeSuspendGetTargetMask: diff --git a/src/util/virsystemd.c b/src/util/virsystemd.c index 93b3f9c..d144d12 100644 --- a/src/util/virsystemd.c +++ b/src/util/virsystemd.c @@ -325,3 +325,62 @@ virSystemdNotifyStartup(void) sd_notify(0, "READY=1"); #endif } + +static int +virSystemdPMSupportTarget(const char *methodName, bool *result) +{ + int ret; + DBusConnection *conn; + DBusMessage *message = NULL; + char *response; + + ret = virDBusIsServiceEnabled("org.freedesktop.login1"); + if (ret < 0) + return ret; + + if ((ret = virDBusIsServiceRegistered("org.freedesktop.login1")) < 0) + return ret; + + if (!(conn = virDBusGetSystemBus())) + return -1; + + ret = -1; + + if (virDBusCallMethod(conn, + &message, + NULL, + "org.freedesktop.login1", + "/org/freedesktop/login1", + "org.freedesktop.login1.Manager", + methodName, + NULL) < 0) + return ret; + + if ((ret = virDBusMessageRead(message, "s", &response)) < 0) + goto cleanup; + + *result = STRNEQ("no", response); + + ret = 0; + + cleanup: + dbus_message_unref(message); + VIR_FREE(response); + + return ret; +} + +int virSystemdCanSuspend(bool *result) +{ + return virSystemdPMSupportTarget("CanSuspend", result); +} + +int virSystemdCanHibernate(bool *result) +{ + return virSystemdPMSupportTarget("CanHibernate", result); +} + +int virSystemdCanHybridSleep(bool *result) +{ + return virSystemdPMSupportTarget("CanHybridSleep", result); +} diff --git a/src/util/virsystemd.h b/src/util/virsystemd.h index 7fed456..491c9b7 100644 --- a/src/util/virsystemd.h +++ b/src/util/virsystemd.h @@ -48,4 +48,10 @@ int virSystemdTerminateMachine(const char *name, void virSystemdNotifyStartup(void); +int virSystemdCanSuspend(bool *result); + +int virSystemdCanHibernate(bool *result); + +int virSystemdCanHybridSleep(bool *result); + #endif /* __VIR_SYSTEMD_H__ */ -- 1.8.4.5

4 5

[libvirt] [PATCH] Remove useless NULL check in virSecurityManagerGenLabel
by Ján Tomko 02 Apr '14

02 Apr '14

Every security driver has domainGenSecurityLabel defined. Coverity complains about a possible leak of seclabel if !sec_managers[i]->drv->domainGenSecurityLabel is true and the seclabel might be overwritten by the next iteration of the loop. --- src/security/security_manager.c | 28 ++++++++++++---------------- 1 file changed, 12 insertions(+), 16 deletions(-) diff --git a/src/security/security_manager.c b/src/security/security_manager.c index d68c7e9..24855db 100644 --- a/src/security/security_manager.c +++ b/src/security/security_manager.c @@ -512,24 +512,20 @@ int virSecurityManagerGenLabel(virSecurityManagerPtr mgr, } } - if (!sec_managers[i]->drv->domainGenSecurityLabel) { - virReportUnsupportedError(); - } else { - /* The seclabel must be added to @vm prior calling domainGenSecurityLabel - * which may require seclabel to be presented already */ - if (generated && - VIR_APPEND_ELEMENT(vm->seclabels, vm->nseclabels, seclabel) < 0) - goto cleanup; - - if (sec_managers[i]->drv->domainGenSecurityLabel(sec_managers[i], vm) < 0) { - if (VIR_DELETE_ELEMENT(vm->seclabels, - vm->nseclabels -1, vm->nseclabels) < 0) - vm->nseclabels--; - goto cleanup; - } + /* The seclabel must be added to @vm prior calling domainGenSecurityLabel + * which may require seclabel to be presented already */ + if (generated && + VIR_APPEND_ELEMENT(vm->seclabels, vm->nseclabels, seclabel) < 0) + goto cleanup; - seclabel = NULL; + if (sec_managers[i]->drv->domainGenSecurityLabel(sec_managers[i], vm) < 0) { + if (VIR_DELETE_ELEMENT(vm->seclabels, + vm->nseclabels -1, vm->nseclabels) < 0) + vm->nseclabels--; + goto cleanup; } + + seclabel = NULL; } ret = 0; -- 1.8.3.2

3 3

[libvirt] [PATCHv2 0/4] Time setting and getting in qemu guests
by Michal Privoznik 02 Apr '14

02 Apr '14

If we'd come up with a struct to interpret the time, it's written in stone and there's nothing we can do about it. Even if we break up the struct into function arguments. However, we can use typed parameters and have the API extensible. For example, I'm implementing seconds granularity only for now. If later somebody feels like he/she wants to use even finer granularity, we can do that by inventing a new typed param. Same goes for timezone, etc. Michal Privoznik (4): Introduce virDomain{Get,Set}Time APIs remote: Implement remote{Get,Set}Time virsh: Expose virDomain{Get,Set}Time qemu: Implement virDomain{Get,Set}Time daemon/remote.c | 50 +++++++++++++++ include/libvirt/libvirt.h.in | 24 ++++++++ src/access/viraccessperm.c | 2 +- src/access/viraccessperm.h | 7 ++- src/driver.h | 14 +++++ src/libvirt.c | 96 +++++++++++++++++++++++++++++ src/libvirt_public.syms | 6 ++ src/qemu/qemu_agent.c | 95 ++++++++++++++++++++++++++++ src/qemu/qemu_agent.h | 6 ++ src/qemu/qemu_driver.c | 139 +++++++++++++++++++++++++++++++++++++++++ src/remote/remote_driver.c | 47 ++++++++++++++ src/remote/remote_protocol.x | 32 +++++++++- src/remote_protocol-structs | 20 ++++++ tools/virsh-domain-monitor.c | 143 +++++++++++++++++++++++++++++++++++++++++++ tools/virsh.pod | 16 +++++ 15 files changed, 694 insertions(+), 3 deletions(-) -- 1.9.0

3 6

Re: [libvirt] RFC: viridentitytest Failure on CentOS 6.5?
by Nehal J Wani 02 Apr '14

02 Apr '14

On Sun, Mar 16, 2014 at 1:28 PM, Nehal J Wani <nehaljw.kkd1(a)gmail.com> wrote: > I followed the following steps to build a clean development version of > libvirt from source. Can't seem to understand what is wrong. (Entire > buildLog has been attached): > > ➜ /tmp cat /etc/issue > CentOS release 6.5 (Final) > Kernel \r on an \m > ➜ /tmp sestatus > SELinux status: enabled > SELinuxfs mount: /selinux > Current mode: enforcing > Mode from config file: enforcing > Policy version: 24 > Policy from config file: targeted > ➜ /tmp git clone git://libvirt.org/libvirt.git > Initialized empty Git repository in /tmp/libvirt/.git/ > remote: Counting objects: 138935, done. > remote: Compressing objects: 100% (21143/21143), done. > remote: Total 138935 (delta 117282), reused 138935 (delta 117282) > Receiving objects: 100% (138935/138935), 137.72 MiB | 559 KiB/s, done. > Resolving deltas: 100% (117282/117282), done. > ➜ /tmp cd libvirt > ➜ libvirt git:(master) ./autogen.sh --system > Running ./configure with --prefix=/usr --sysconfdir=/etc > --localstatedir=/var --libdir=/usr/lib64 > .. > .. > Now type 'make' to compile libvirt. > ➜ libvirt git:(master) make > .. > .. > make[1]: Leaving directory `/tmp/libvirt' > ➜ libvirt git:(master) ✗ cd tests > ➜ tests git:(master) ✗ ./viridentitytest > /tmp/libvirt/tests/.libs/libsecurityselinuxhelper.so: No such file or directory > ➜ tests git:(master) ✗ > > > -- > Nehal J Wani Ping! -- Nehal J Wani

3 4

[libvirt] [PATCH] qemu: cleanup error checking on agent replies
by Martin Kletzander 02 Apr '14

02 Apr '14

On all the places where qemuAgentComand() was called, we did a check for errors in the reply. Unfortunately, some of the places called qemuAgentCheckError() without checking for non-null reply which might have resulted in a crash. So this patch makes the error-checking part of qemuAgentCommand() itself, which: a) makes it look better, b) makes the check mandatory and, most importantly, c) checks for the errors if and only if it is appropriate. This actually fixes a potential crashers when qemuAgentComand() returned 0, but reply was NULL. Having said that, it *should* fix the following bug: https://bugzilla.redhat.com/show_bug.cgi?id=1058149 Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com> --- src/qemu/qemu_agent.c | 27 +++++++-------------------- 1 file changed, 7 insertions(+), 20 deletions(-) diff --git a/src/qemu/qemu_agent.c b/src/qemu/qemu_agent.c index 27af1be..4c83d7d 100644 --- a/src/qemu/qemu_agent.c +++ b/src/qemu/qemu_agent.c @@ -117,6 +117,8 @@ struct _qemuAgent { qemuAgentEvent await_event; }; +static int qemuAgentCheckError(virJSONValuePtr cmd, virJSONValuePtr reply); + static virClassPtr qemuAgentClass; static void qemuAgentDispose(void *obj); @@ -1009,6 +1011,7 @@ qemuAgentCommand(qemuAgentPtr mon, } } else { *reply = msg.rxObject; + ret = qemuAgentCheckError(cmd, *reply); } } @@ -1277,9 +1280,6 @@ int qemuAgentShutdown(qemuAgentPtr mon, ret = qemuAgentCommand(mon, cmd, &reply, VIR_DOMAIN_QEMU_AGENT_COMMAND_BLOCK); - if (reply && ret == 0) - ret = qemuAgentCheckError(cmd, reply); - virJSONValueFree(cmd); virJSONValueFree(reply); return ret; @@ -1308,8 +1308,7 @@ int qemuAgentFSFreeze(qemuAgentPtr mon) return -1; if (qemuAgentCommand(mon, cmd, &reply, - VIR_DOMAIN_QEMU_AGENT_COMMAND_BLOCK) < 0 || - qemuAgentCheckError(cmd, reply) < 0) + VIR_DOMAIN_QEMU_AGENT_COMMAND_BLOCK) < 0) goto cleanup; if (virJSONValueObjectGetNumberInt(reply, "return", &ret) < 0) { @@ -1346,8 +1345,7 @@ int qemuAgentFSThaw(qemuAgentPtr mon) return -1; if (qemuAgentCommand(mon, cmd, &reply, - VIR_DOMAIN_QEMU_AGENT_COMMAND_BLOCK) < 0 || - qemuAgentCheckError(cmd, reply) < 0) + VIR_DOMAIN_QEMU_AGENT_COMMAND_BLOCK) < 0) goto cleanup; if (virJSONValueObjectGetNumberInt(reply, "return", &ret) < 0) { @@ -1386,9 +1384,6 @@ qemuAgentSuspend(qemuAgentPtr mon, ret = qemuAgentCommand(mon, cmd, &reply, VIR_DOMAIN_QEMU_AGENT_COMMAND_BLOCK); - if (reply && ret == 0) - ret = qemuAgentCheckError(cmd, reply); - virJSONValueFree(cmd); virJSONValueFree(reply); return ret; @@ -1419,9 +1414,6 @@ qemuAgentArbitraryCommand(qemuAgentPtr mon, if ((ret = qemuAgentCommand(mon, cmd, &reply, timeout)) < 0) goto cleanup; - if ((ret = qemuAgentCheckError(cmd, reply)) < 0) - goto cleanup; - if (!(*result = virJSONValueToString(reply, false))) ret = -1; @@ -1449,9 +1441,6 @@ qemuAgentFSTrim(qemuAgentPtr mon, ret = qemuAgentCommand(mon, cmd, &reply, VIR_DOMAIN_QEMU_AGENT_COMMAND_BLOCK); - if (reply && ret == 0) - ret = qemuAgentCheckError(cmd, reply); - virJSONValueFree(cmd); virJSONValueFree(reply); return ret; @@ -1472,8 +1461,7 @@ qemuAgentGetVCPUs(qemuAgentPtr mon, return -1; if (qemuAgentCommand(mon, cmd, &reply, - VIR_DOMAIN_QEMU_AGENT_COMMAND_BLOCK) < 0 || - qemuAgentCheckError(cmd, reply) < 0) + VIR_DOMAIN_QEMU_AGENT_COMMAND_BLOCK) < 0) goto cleanup; if (!(data = virJSONValueObjectGet(reply, "return"))) { @@ -1581,8 +1569,7 @@ qemuAgentSetVCPUs(qemuAgentPtr mon, cpus = NULL; if (qemuAgentCommand(mon, cmd, &reply, - VIR_DOMAIN_QEMU_AGENT_COMMAND_BLOCK) < 0 || - qemuAgentCheckError(cmd, reply) < 0) + VIR_DOMAIN_QEMU_AGENT_COMMAND_BLOCK) < 0) goto cleanup; if (virJSONValueObjectGetNumberInt(reply, "return", &ret) < 0) { -- 1.9.1

2 3

[libvirt] [PATCH] virsh: man: delete the unexpected character in snapshot-list
by Shanzhi Yu 02 Apr '14

02 Apr '14

--- tools/virsh.pod | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tools/virsh.pod b/tools/virsh.pod index ba2da20..98d891a 100644 --- a/tools/virsh.pod +++ b/tools/virsh.pod @@ -3162,7 +3162,7 @@ with I<--current>. =item B<snapshot-list> I<domain> [I<--metadata>] [I<--no-metadata>] [{I<--parent> | I<--roots> | [{I<--tree> | I<--name>}]}] [{[I<--from>] B<snapshot> | I<--current>} [I<--descendants>]] -[I<--leaves>] [I<--no-leaves>] p[I<--inactive>] [I<--active>] +[I<--leaves>] [I<--no-leaves>] [I<--inactive>] [I<--active>] [I<--disk-only>] [I<--internal>] [I<--external>] List all of the available snapshots for the given domain, defaulting -- 1.9.0

3 2

[libvirt] [BUG] xenInotify/keep-alive double-free
by Philipp Hahn 02 Apr '14

02 Apr '14

Hello, There seems to be a double-free bug in libvirt; I've checked libvirt 0.9.12 and 1.2.3: With 0.9.12 xenUnifiedOpen() passes a pointer of "conn" to xenInotifyOpen(), which passes it to virEventAddHandle(...opaque=conn...). After successfully defining a new domain inotify picks up the newly created directory and generates an event, which is processes when virsh already dropped its last public reference to the domain and subsequently already freed the connection. Since *conn is not zero-filled, the data is still valid and usable, but virsh the terminates on double-freeing some internal data. I can reproduce by it doing: virsh undefine $DOM virsh -c xen:// define $DOM.xml For 1.2.3 something similar seems to happen with the keep-alive: $ grep unref ~/BUG/31032_virsh-define-segv.log virUnrefDomain:276 : unref domain 0x7f4ec4003fe0 ucs32-64-segv 1 virReleaseDomain:246 : unref connection 0x917650 2 virUnrefDomain:276 : unref domain 0x934460 ucs32-64-segv 1 virReleaseDomain:246 : unref connection 0x917650 2 virUnrefConnect:145 : unref connection 0x917650 1 virUnrefDomain:276 : unref domain 0x7f4ec4004060 ucs32-64-segv 1 virReleaseDomain:246 : unref connection 0x917650 1 Notice that there are two lines for "unref connection ... 1"! My gut feeling is that libvirt should also increment the reference counter for internal references to delay freeing still used data and add a second counter to track external references, which is used to start closing down things. Comments and ideas welcomed. We're tracking this as <https://forge.univention.org/bugzilla/show_bug.cgi?id=31032> Sincerely Philipp -- Philipp Hahn Open Source Software Engineer Univention GmbH be open. Mary-Somerville-Str. 1 D-28359 Bremen Tel.: +49 421 22232-0 Fax : +49 421 22232-99 hahn(a)univention.de http://www.univention.de/ Geschäftsführer: Peter H. Ganten HRB 20755 Amtsgericht Bremen Steuer-Nr.: 71-597-02876

1 0

[libvirt] [bug?] unix sockets opened via chardev devices not being closed on shutdown
by Chris Friesen 01 Apr '14

01 Apr '14

I have a case where I'm creating a virtio channel between the host and guest using something like this: <channel type="unix"> <source mode="bind" path="/path/in/host/instance_name"/> <target type="virtio" name="name_in_guest"/> </channel> When qemu is started up this gets created as expected, but when qemu is shut down the unix socket is left in the filesystem. It seems to me that libvirt should be deleting this unix socket the same way that it deletes the monitor socket in qemuProcessStop(). Anyone else trying to delete it is going to be subject to race conditions since they can't know whether or not a virtual machine has been (re)created that wants to use the same socket path. Chris

1 0