August 2009 - Devel - libvirt List Archives

[libvirt] [PATCH 00/08] More misc device assignment fixes
by Mark McLoughlin 17 Aug '09

17 Aug '09

Hey, I had a little change of heart about the patches committed last week to allow pciResetDevice() reset multiple devices. The callback to check whether a device is active is much to cumbersome, instead we should maintain a list of active devices. Patch 2 reverts two commits from last week and patch 7 and 8 replace them with something much better. The reset of the patches are a bunch of fairly minor fixes. Cheers, Mark.

2 15

[libvirt] [PATCH] Cosmetic change to 'virsh nodedev-list --tree' output
by Mark McLoughlin 17 Aug '09

17 Aug '09

Maybe it's just me, but I try to select an item from the tree using double-click and get annoyed when "+-" gets included in the selection. * src/virsh.c: add a space between "+-" and the node device name in 'virsh nodedev-list --tree' --- src/virsh.c | 4 +++- 1 files changed, 3 insertions(+), 1 deletions(-) diff --git a/src/virsh.c b/src/virsh.c index 94c3c4e..2d0cf81 100644 --- a/src/virsh.c +++ b/src/virsh.c @@ -5370,6 +5370,8 @@ cmdNodeListDevicesPrint(vshControl *ctl, if (depth && depth < MAX_DEPTH) { indentBuf[indentIdx] = '+'; indentBuf[indentIdx+1] = '-'; + indentBuf[indentIdx+2] = ' '; + indentBuf[indentIdx+3] = '\0'; } /* Print this device */ @@ -5398,7 +5400,7 @@ cmdNodeListDevicesPrint(vshControl *ctl, /* If there is a child device, then print another blank line */ if (nextlastdev != -1) { vshPrint(ctl, "%s", indentBuf); - vshPrint(ctl, " |\n"); + vshPrint(ctl, " |\n"); } /* Finally print all children */ -- 1.6.2.5

2 3

[libvirt] XenStore fix
by Jonas Eriksson 17 Aug '09

17 Aug '09

Hi, I have been examining a bug where libvirtd (and virsh) does not show all virtual machines on a xen host. This proved to be because of this program flow: 1. virConnectNumOfDomains -> .. -> xenUnifiedNumOfDomains -> xenHypervisorNumOfDomains => 3 2. virConnectListDomains(max=3) -> .. -> xenUnifiedListDomains(max=3) -> xenStoreNumOfDomains(max=3) => { 0, 2, 7 } The domain with ID 2 is then removed when it is discovered that it is not a running domain, which leads to this: xenhost# xm list Name ID Mem VCPUs State Time(s) Domain-0 0 14970 2 r----- 2544.7 vm1 7 512 1 -b---- 2191.7 vm4 512 1 28.0 vm5 12 512 1 -b---- 467.1 vm6 512 1 0.0 vm7 512 1 482.4 xenhost# virsh list Id Name State ---------------------------------- 0 Domain-0 running 7 vm1 idle xenhost# But where does "2" come from? If we check all "directories" in /local/domain which is queried by the xenstore driver, it is apparent that xenstore is not properly cleaned. We find the sequence {0, 2, 7} as the first entries: xenhost# xenstore ls /local/domain |grep '^[^ ]' 0 = "" 2 = "" 7 = "" 9 = "" 10 = "" 11 = "" 12 = "" xenhost# This patch checks that the path found in /local/domain/<domid>/vm exists in xenstore before adding the domid to the return list. The same thing is done for xenStoreNumOfDomains. I use SLES11 with Xen 3.3.1_18546_12-3.1. /Jonas

4 16

[libvirt] libvirt socket closed unexpectedly (code=39) & libvirt: Broken pipe (code=38)
by dave c 17 Aug '09

17 Aug '09

Hello everyone, I'm trying to get eucalyptus v1.5.2 working on debian lenny 64bit, the relevant info I can give you is as follows: --- libvirtd conf file ---- 01:/etc/libvirt# cat libvirtd.conf|grep -v '#' unix_sock_group = "libvirt" unix_sock_rw_perms = "0770" auth_unix_ro = "none" auth_unix_rw = "none" --------------------- version of libvir installed on my server (cloud controller/cluster controller): 01:/etc/libvirt# dpkg -l|grep libvirt ii libvirt-bin 0.4.6-10 the programs for the libvirt library ii libvirt-dev 0.4.6-10 development files for the libvirt library ii libvirt-doc 0.4.6-10 documentation for the libvirt library ii libvirt0 0.4.6-10 library for interfacing with different virtualization systems version installed on my node: 02:~# dpkg -l |grep libvirt ii libvirt-bin 0.4.6-10 the programs for the libvirt library ii libvirt0 0.4.6-10 library for interfacing with different virtualization systems ================ error in my nc..log (eucalyptus node log): [Sat Aug 15 02:24:58 2009][005115][EUCADEBUG ] system_output(): [//usr/lib/eucalyptus/euca_rootwrap //usr/share/eucalyptus/gen_kvm_libvirt_xml --ramdisk --ephemeral] [Sat Aug 15 02:24:58 2009][005115][EUCAERROR ] libvirt: socket closed unexpectedly (code=39) [Sat Aug 15 02:24:58 2009][005115][EUCAERROR ] libvirt: Broken pipe (code=38) basically when i start up an instance, it starts up and then terminates within 1 min I've already posted on the eucalyptus forums, i thought the libvir-list might be able to assist as well thoughts / ideas ? anything else I can share ?

2 1

[libvirt] remote_protocol.c not compiling
by Kenneth Nagin 17 Aug '09

17 Aug '09

I am experimenting with some new APIs in Libvirt 0.6.5 and added the changes to remote_protocol.x . I'm getting a make error of undefined reference in ./.libs/libvirt.so to the xdr arguments. I think it is related to the fact that a new remote_protocol.c is not being generated. I've tried the obvious "make clean" and "make" and "make -C qemud remote_protocol.c." But it has no affect. [nagin@gorky trunk]$ make -C qemud remote_protocol.c make: Entering directory `/gpfs/reservoir/nagin/workspace/LIBVIRT/LIBVIRT_0_6_5/trunk/qemud' make: Nothing to be done for `remote_protocol.c'. make: Leaving directory `/gpfs/reservoir/nagin/workspace/LIBVIRT/LIBVIRT_0_6_5/trunk/qemud' Any suggestions? Kenneth Nagin

2 2

[libvirt] [PATCH] Compressed save image format for Qemu.
by Chris Lalancette 17 Aug '09

17 Aug '09

Implement a compressed save image format for qemu. While ideally we would have the choice between compressed/non-compressed available to the libvirt API, unfortunately there is no "flags" parameter to the virDomainSave() API. Therefore, implement this as a qemu.conf option. Both gzip and bzip2 are implemented, and it should be very easy to implement additional compression methods. One open question is if/how we should detect the gzip and bzip2 binaries. One way to do it is to do compile-time setting of the paths (via configure.in), but that doesn't seem like a great thing to do. Another solution (my preferred solution) is not to detect at all; when we go to run the commands that need them, if they aren't available, or aren't available in one of the standard paths, then we'll fail. Maybe somebody else has another option or opinion, though. In the future, we'll have a more robust (managed) save/restore API, at which time we can expose this functionality properly in the API. V2: get rid of redundant dd command and just use >> to append data. V3: Add back the missing pieces for the enum and bumping the save version. V4: Make the compressed field in the save_header an int. Implement LZMA compression. Signed-off-by: Chris Lalancette <clalance(a)redhat.com> --- src/qemu.conf | 10 ++++++ src/qemu_conf.c | 11 ++++++ src/qemu_conf.h | 2 + src/qemu_driver.c | 93 +++++++++++++++++++++++++++++++++++++++++++++++++---- 4 files changed, 109 insertions(+), 7 deletions(-) diff --git a/src/qemu.conf b/src/qemu.conf index 653f487..1f10b43 100644 --- a/src/qemu.conf +++ b/src/qemu.conf @@ -129,3 +129,13 @@ # "/dev/ptmx", "/dev/kvm", "/dev/kqemu", # "/dev/rtc", "/dev/hpet", "/dev/net/tun", #] + +# The default format for Qemu/KVM guest save images is raw; that is, the +# memory from the domain is dumped out directly to a file. If you have +# guests with a large amount of memory, however, this can take up quite +# a bit of space. If you would like to compress the images while they +# are being saved to disk, you can also set "gzip", "bzip2", or "lzma" +# for save_image_format. Note that this means you slow down the +# process of saving a domain in order to save disk space. +# +# save_image_format = "raw" diff --git a/src/qemu_conf.c b/src/qemu_conf.c index 7ca5a15..ed87e13 100644 --- a/src/qemu_conf.c +++ b/src/qemu_conf.c @@ -280,6 +280,17 @@ int qemudLoadDriverConfig(struct qemud_driver *driver, driver->cgroupDeviceACL[i] = NULL; } + p = virConfGetValue (conf, "save_image_format"); + CHECK_TYPE ("save_image_format", VIR_CONF_STRING); + if (p && p->str) { + VIR_FREE(driver->saveImageFormat); + if (!(driver->saveImageFormat = strdup(p->str))) { + virReportOOMError(NULL); + virConfFree(conf); + return -1; + } + } + virConfFree (conf); return 0; } diff --git a/src/qemu_conf.h b/src/qemu_conf.h index 8f4ef6a..e34baab 100644 --- a/src/qemu_conf.h +++ b/src/qemu_conf.h @@ -111,6 +111,8 @@ struct qemud_driver { char *securityDriverName; virSecurityDriverPtr securityDriver; + + char *saveImageFormat; }; diff --git a/src/qemu_driver.c b/src/qemu_driver.c index 20906ef..3fd153d 100644 --- a/src/qemu_driver.c +++ b/src/qemu_driver.c @@ -3411,18 +3411,27 @@ static char *qemudEscapeShellArg(const char *in) } #define QEMUD_SAVE_MAGIC "LibvirtQemudSave" -#define QEMUD_SAVE_VERSION 1 +#define QEMUD_SAVE_VERSION 2 + +enum qemud_save_formats { + QEMUD_SAVE_FORMAT_RAW, + QEMUD_SAVE_FORMAT_GZIP, + QEMUD_SAVE_FORMAT_BZIP2, + QEMUD_SAVE_FORMAT_LZMA, +}; struct qemud_save_header { char magic[sizeof(QEMUD_SAVE_MAGIC)-1]; int version; int xml_len; int was_running; - int unused[16]; + int compressed; + int unused[15]; }; static int qemudDomainSave(virDomainPtr dom, - const char *path) { + const char *path) +{ struct qemud_driver *driver = dom->conn->privateData; virDomainObjPtr vm; char *command = NULL; @@ -3433,11 +3442,28 @@ static int qemudDomainSave(virDomainPtr dom, struct qemud_save_header header; int ret = -1; virDomainEventPtr event = NULL; + int internalret; memset(&header, 0, sizeof(header)); memcpy(header.magic, QEMUD_SAVE_MAGIC, sizeof(header.magic)); header.version = QEMUD_SAVE_VERSION; + if (driver->saveImageFormat == NULL) + header.compressed = QEMUD_SAVE_FORMAT_RAW; + else if (STREQ(driver->saveImageFormat, "raw")) + header.compressed = QEMUD_SAVE_FORMAT_RAW; + else if (STREQ(driver->saveImageFormat, "gzip")) + header.compressed = QEMUD_SAVE_FORMAT_GZIP; + else if (STREQ(driver->saveImageFormat, "bzip2")) + header.compressed = QEMUD_SAVE_FORMAT_BZIP2; + else if (STREQ(driver->saveImageFormat, "lzma")) + header.compressed = QEMUD_SAVE_FORMAT_LZMA; + else { + qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED, + "%s", _("Invalid save image format specified in configuration file")); + return -1; + } + qemuDriverLock(driver); vm = virDomainFindByUUID(&driver->domains, dom->uuid); @@ -3510,11 +3536,28 @@ static int qemudDomainSave(virDomainPtr dom, virReportOOMError(dom->conn); goto cleanup; } - if (virAsprintf(&command, "migrate \"exec:" - "dd of='%s' oflag=append conv=notrunc 2>/dev/null" - "\"", safe_path) == -1) { + + if (header.compressed == QEMUD_SAVE_FORMAT_RAW) + internalret = virAsprintf(&command, "migrate \"exec:" + "dd of='%s' oflag=append conv=notrunc 2>/dev/null" + "\"", safe_path); + else if (header.compressed == QEMUD_SAVE_FORMAT_GZIP) + internalret = virAsprintf(&command, "migrate \"exec:" + "gzip -c >> '%s' 2>/dev/null\"", safe_path); + else if (header.compressed == QEMUD_SAVE_FORMAT_BZIP2) + internalret = virAsprintf(&command, "migrate \"exec:" + "bzip2 -c >> '%s' 2>/dev/null\"", safe_path); + else if (header.compressed == QEMUD_SAVE_FORMAT_LZMA) + internalret = virAsprintf(&command, "migrate \"exec:" + "lzma -c >> '%s' 2>/dev/null\"", safe_path); + else { + qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR, + _("Invalid compress format %d"), + header.compressed); + goto cleanup; + } + if (internalret < 0) { virReportOOMError(dom->conn); - command = NULL; goto cleanup; } @@ -4035,6 +4078,9 @@ static int qemudDomainRestore(virConnectPtr conn, char *xml = NULL; struct qemud_save_header header; virDomainEventPtr event = NULL; + int intermediatefd = -1; + pid_t intermediate_pid = -1; + int childstat; qemuDriverLock(driver); /* Verify the header and read the XML */ @@ -4124,8 +4170,41 @@ static int qemudDomainRestore(virConnectPtr conn, } def = NULL; + if (header.version == 2) { + const char *intermediate_argv[3] = { NULL, "-dc", NULL }; + if (header.compressed == QEMUD_SAVE_FORMAT_GZIP) + intermediate_argv[0] = "gzip"; + else if (header.compressed == QEMUD_SAVE_FORMAT_BZIP2) + intermediate_argv[0] = "bzip2"; + else if (header.compressed == QEMUD_SAVE_FORMAT_LZMA) + intermediate_argv[0] = "lzma"; + else if (header.compressed != QEMUD_SAVE_FORMAT_RAW) { + qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED, + _("Unknown compressed save format %d"), + header.compressed); + goto cleanup; + } + if (intermediate_argv[0] != NULL) { + intermediatefd = fd; + fd = -1; + if (virExec(conn, intermediate_argv, NULL, NULL, + &intermediate_pid, intermediatefd, &fd, NULL, 0) < 0) { + qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR, + _("Failed to start decompression binary %s"), + intermediate_argv[0]); + goto cleanup; + } + } + } /* Set the migration source and start it up. */ ret = qemudStartVMDaemon(conn, driver, vm, "stdio", fd); + if (intermediate_pid != -1) { + /* Wait for intermediate process to exit */ + while (waitpid(intermediate_pid, &childstat, 0) == -1 && + errno == EINTR); + } + if (intermediatefd != -1) + close(intermediatefd); close(fd); fd = -1; if (ret < 0) { -- 1.6.0.6

2 2

[libvirt] [RFC] Interface for disk hotadd/remove
by Wolfgang Mauerer 14 Aug '09

14 Aug '09

Hi, I'm currently interested in implementing hard disk hot-add and -remove support for qemu (as opposed to controller-based hotplugging), and this brings up the question how to best support this feature in libvirt. Many SCSI-Controllers in real machines, for instance, allow to add and remove disks (without adding or removing the controller itself) while the system is up and running, so it would be nice to emulate this in a virtual machine. I'm focusing on qemu on the backend side, but the problem is not related to this particular backend. Rather, the question is how to integrate such a feature best into libvirt. Before implementing the functionality, it would be great to hear the community's opinion which route to take wrt. the interface. Essentially, I can see two options: - Naturally, there are virDomain{At,De}tachDevice, but the pair implements drive-hotadding via adding a new controller with an attached hard disk to the system. By extending the XML description of the drive with a parameter that specifies whether controller- or disk-based hotplugging is to be performed, it would be possible to implement the desired functionality, whilst preserving compatibility with existing semantics. Removing the drive would then require another new parameter in the XML description to identify the drive on the controller, which does not really prettify the thing. - Extend the API with a new method (for instance virDomainDiskAttach) that takes a hard disk description, a controller identifier, and a parameter that identifies the disk on the controller. - (Theoretically, it would also be possible to implement media exchange for hard disks in qemu and re-use the media exchange infrastructure already present in libvirt for CD-ROMs, but since this possibility comes to use on real hardware only very occasionally, guest operating systems are typically not really prepared to handle this well) My preference would be to go for option 2, that is, implement a new API method. Would there be any obstacles against adding such a patch to mainline? Or is anyone already working on similar functionality? Or can this be done in a much simpler way I've missed? If not, then I'd send patches for more detailed review before long. Thanks, Wolfgang

2 3

[libvirt] Share storage using iscsi
by Łukasz Mierzwa 14 Aug '09

14 Aug '09

Hi, I'm trying to setup pool of machines (nodes) for virtual machines hosting and I got few question about shared storage. My main requirements are: 1. central management - I've got simple python app that stores information about all virtual machines and all nodes, this app needs to be able to manage volumes using libvirt API, so I need libvirt volume pools 2. live migration - I got shared storage with HA, I want to use it also for live migration in case one of nodes is dying or if I want to do some load balancing Right now I'm thinking about 2 machines with disks synchronized using drdb, both acting as a identical iscsi targets, iscsi HA will be provided by heartbeat. So I will end up with virtual IP pointing to working iscsi target, drdb should keep storage is sync. But: 1. I can't just use single iscsi LUN and export it as libvirt storage pool to each node, because no pool type would work that way, right? 2. http://libvirt.org/storage.html section "iSCSI volume pools" says: "Volumes must be pre-allocated on the iSCSI server, and cannot be created via the libvirt APIs." So even if I got one LUN per node and set it as iscsi volume pool I would need to create each volume on iscsi target. Libvirt can't manage volumes in such pool, it can only assign already created volumes to virtual machines, right? 3. So maybe my storage could be setup as LVM volume group and this lvm group would be managed as libvirt lvm volume pool on master (from heartbeat POV) iscsi target. I would create one logical volume per virtual machine, export this volume as a separate iscsi LUN, and use this LUN as iscsi volume for virtual machine. To create new virtual machine I would: a) create lvm volume on iscsi target using libvirt (so I don't have to Does it makes any sense? Are there better ways to *manage* volumes for virtual machines using iscsi? Thanks for any tips. Łukasz Mierzwa

1 0

[libvirt] [PATCH 1/2]: VirtualBox: Updated vboxNetworkCreateXML() and vboxNetworkDefineXML()
by Pritesh Kothari 13 Aug '09

13 Aug '09

Hi All, I have made some changes to the functions vboxNetworkCreateXML(), vboxNetworkDefineXML(), vboxNetworkUndefine() and vboxNetworkDestroy() to handle multiple host only interfaces as multiple host only interfaces are supported by VirtualBox 3.0 and greater. The patch's are as below: PATCH 1/2: Merged vboxNetworkCreateXML() and vboxNetworkDefineXML() and added code to handle multiple hostonly interfaces. PATCH 2/2: Merged vboxNetworkUndefine() and vboxNetworkDestroy() and added code to handle multiple hostonly interfaces. Regards, Pritesh

2 1

[libvirt] Connection refcounting fix?
by Chris Lalancette 13 Aug '09

13 Aug '09

DanB, is this the sort of idea you had to fix up the connection refcounting?

2 2