- Devel - libvirt List Archives

[Libvir] "virsh list" command of libvirt consumes a lot of CPU in the domain-0
by jean-paul.pigache＠bull.net 12 Mar '08

12 Mar '08

Hi all, I know that this is not a libvirt issue but this badly impacts libvirt usage. Is anyone aware of any status on this issue ? Daniel ? Here is some history I could get from the libvirt mailing list : * October 12, 2006 (Daniel Berrange). I've been trying to track down just why talking to XenD is resulting in so much CPU time being comsumed by both xend & xenstored. As a test case, I'm running 'virsh dominfo demo' which results in a single HTTP request to Xend to fetch domain info, eg 'GET /xend/domains/demo' Run this in a tight loop & I'll see xenstored taking > 50% CPU, and XenD taking another 11% PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 2647 root 16 0 6188 840 464 R 52 0.0 0:55.04 xenstored 11600 root 18 0 259m 7568 1516 S 11 0.2 0:04.53 python Its not surprising that xend is consuming time since we are making many requests per second, but for an operation which is only doing reads it having so much time attributed to xenstored seems very excessive. So I ran oprofile & collected some data about xenstored: CPU: AMD64 processors, speed 2211.33 MHz (estimated) Counted CPU_CLK_UNHALTED events (Cycles outside of halt state) with a unit mask of 0x00 (No unit mask) count 100000 samples % image name symbol name 347226 45.9445 ext3 (no symbols) 264664 35.0200 jbd (no symbols) 31778 4.2048 libc-2.5.so memset 10763 1.4241 xenstored main 8884 1.1755 libc-2.5.so _int_malloc 7053 0.9332 libc-2.5.so vfprintf 4264 0.5642 xenstored initialize_set So almost 80% of xenstored's CPU time is attributed to ext3 & journalling modules, suggesting xenstored is doing alot of disk I/O. strace()'ing the xenstored process shows the only file it is opening is: # strace -p 2647 -e trace=open,rename,unlink Process 2647 attached - interrupt to quit open("/var/lib/xenstored/tdb.0x62aa80", O_WRONLY|O_CREAT|O_TRUNC, 0640) = 13 open("/var/lib/xenstored/tdb.0x62aa80", O_RDWR) = 15 rename("/var/lib/xenstored/tdb.0x62aa80", "/var/lib/xenstored/tdb") = 0 unlink("/var/lib/xenstored/tdb.0x62aa80") = -1 ENOENT (No such file or directory) open("/var/lib/xenstored/tdb.0x62b2b0", O_WRONLY|O_CREAT|O_TRUNC, 0640) = 13 open("/var/lib/xenstored/tdb.0x62b2b0", O_RDWR) = 14 rename("/var/lib/xenstored/tdb.0x62b2b0", "/var/lib/xenstored/tdb") = 0 unlink("/var/lib/xenstored/tdb.0x62b2b0") = -1 ENOENT (No such file or directory) ... So basically it is repeatedly copying its persistent TBD database over and over again. The TDB on this system is 128 KB in size and each individual HTTP GET on /xend/domain/demo is resulting in 16 copies being made. Do the maths - 128 * 16 == 2 MB of reads, and 2 MB of writes - for a single read in XenD. Now if I monitor the status of 20 domains, once per second that's causing 40 MB of writes & 40 MB of reads every second which is utterly ridiculous & completely non scalable for enterprise deployment :-( There's two problems I see here: 1. Why the need for xenstored to be doing any of this I/O in the first place? If the DB needs to be kept on disk at all, it really needs to have a much saner update/transactional model to only update bits which actually change, rather than re-creating the entire DB on every transaction. But it strikes me that the DB could potentially be kept entirely in memory removing the disk I/O completely. Sure yyou wouldn't be able to restart the daemon then, but even today you can't restart xenstored & expect things to still be working. 2. Why does XenD create sooo many transactions in XenStored for a read op ? Having instrumented Xend it sems that the root cause of the problem is the xen.xend.xenstore.xstransact class. This alllows one to start a transaction, do a bunch of reads/writes & then commit the transaction. At the same time though it has a bunch of static 'convenience' methods for read & write which will implicitly start & commit a transaction. Well 90% of the code in XenD seems to be using these 'convenience' methods instead of explicitly starting a transaction to cover a piece of work - the result is a simple GET causes 16 transactions ....and an 'xm create' results in 80 transactions. These convenience methods are utterly destroying performance. Clearly we can't address these for 3.0.3, but I think both of these areas need serious work in 3.0.4 if we want a scalable control plane in Dom0. Fixing the XenD bit looks particularly hard because any single method using the convenience xenstored read functions can be called under many different contexts, so of which needs transactions, others which don't. It ought to be possible to trace back all the calls & make it possible to pass explicit xstransct objects into all calls & then kill off the convenience methods. * Answer, same day (October 12, 2006) Yes, xenstored is very simple minded in many respects. We will certainly be improving this during 3.0.4 development -- I think we can get the costs down very significantly for commonplace operations without enormous effort. * Avril 25, 2007 (Daniel Berrange) > Xen 3.0.3 has a serious performance bug > (see http://lists.xensource.com/archives/html/xen-devel/2006-10/msg00487.html) > This bug is fixed in Xen 3.0.4 No it isn't. The performance bug is actually at least x2 worse in Xen 3.0.4

2 1

[Libvir] Snapshots for Xen and KVM
by Gabriel Kaufmann 12 Mar '08

12 Mar '08

Hi, Does anyone knows if Xen and KVM support snapshots (regardless if Libvirt provides the API for them). If yes, will libvirt add API for it? Best Regards, Gabriel Kaufmann Software Engineer Gabriel.Kaufmann(a)ericom.com<mailto:Gabriel.Kaufmann@ericom.com> Ericom Software Tel (Dir): +972 2 591 1700 Ext 754 Tel (Main): +972 2 591 1700 http://www.ericom.com Access Done Right Empower Enterprise-Wide Access to Microsoft(r) Terminal Server, Virtual Desktops and Legacy Applications

5 4

Re: [Libvir] virsh start problem + patch
by Daniel P. Berrange 11 Mar '08

11 Mar '08

On Tue, Mar 11, 2008 at 06:57:33AM +0100, Toth Istvan wrote: > # virsh list --all > Id Name State > ---------------------------------- > 0 Domain-0 running > - windows-xen shut off > - windows-xen2 no state > > # xm list > Name ID Mem VCPUs State > Time(s) > Domain-0 0 3489 4 r----- > 853.6 > windows-xen 1 512 0 > -b-s-d 39.4 > windows-xen2 4 512 0 > ------ 22.5 > > The windows-xen, and windows-xen2 domains were installed the very same > way, except I've had a Dom0 reboot since I've installed windows-xen, so > xen has had the opportunity to sort it's state out, whil the > windows-xen2 domain is a fresh install. Starting and stopping (by xm) a > freshly installed windows hvm domain does not sort out the state, only a > Dom0 reboot (or a xend restart) does. > > I have attached the output of xm list --long command. This shows that it is all XenD related - inactive domains should *not* have a 'domid' or 'state' field set in the 'xm list --long' output. In both your windows-xen & windows-xen2 VM, these fields are incorrectly present. > This problem indeed does look like a Xend bug, but it turns out that it > does not actually affect virsh. (It does affect virt-manager, as I've > written to the other list) > The other problem is I am pretty sure, a virsh logic bug, and is > independent of the first one. No, it is again a Xen bug. Inactive domains are *mandated* to have a domain ID of -1. They do not exist in the hypervisor anymore and thus have no domain ID associated with them. If XenD is not clearing the domain ID properly this is a XenD problem. That all said, we may need to workaround this XenD brokenness in the libirt XenD driver. Working around it in virsh is the wrong place. > and I get this: > > ./virsh start windows-xen > <warnings> > error: Domain is already active > virDomainGetID(dom) returns 1 > > ./virsh start windows-xen2 > <warnings> > error: Domain is already active > virDomainGetID(dom) returns 3 > > so virDomainGetID() does not return -1, but returns the actual xen > domain id of the managed, but inactive xen domain, and I believe this is > what it should do, > as it's job is not to tell us about the state of the domain, but to tell > the id of the domain, regardles of its state. No. Inactive domains are mandated to have an ID of -1. Dan. -- |: Red Hat, Engineering, Boston -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 :|

1 0

Re: [Libvir] Accessing qemu from windows
by Richard W.M. Jones 11 Mar '08

11 Mar '08

On Tue, Mar 11, 2008 at 06:00:31PM +0200, Gabriel Kaufmann wrote: > I'm sure it actually makes a remote TCP connection since on the server side I get the error: > 'libvir: error : could not connect to qemu:///system?' It actually prints this out on the server side? Can you run the server with the --verbose option, you'll get a lot more detail about what is happening. Rich. -- Richard Jones, Emerging Technologies, Red Hat http://et.redhat.com/~rjones virt-p2v converts physical machines to virtual machines. Boot with a live CD or over the network (PXE) and turn machines into Xen guests. http://et.redhat.com/~rjones/virt-p2v

1 0

[Libvir] Accessing qemu from windows
by Gabriel Kaufmann 11 Mar '08

11 Mar '08

Hi, I'm running virsh from linux machine with the URI qemu+tcp://<user>:<password>@<ip>/system >From a linux machine (not the same that runs the libvirtd deamon) with libvirt 0.3.0, this URI allows me to connect. In the other hand, using the Virsh.exe (that comes with ocaml-libvirt-0.4.0.1.exe<http://libvirt.org/sources/ocaml/ocaml-libvirt-0.4.0.1.exe> ) on Windows XP, I receive the error 'Virtualization error: libvirt: VIR_ERROR_NO_CONNECT: VIR_FROM_NONE: could not connect to qemu:///system?' Is it a known bug, or is it configuration problem on my server? Best Regards, Gabriel Kaufmann Software Engineer Gabriel.Kaufmann(a)ericom.com<mailto:Gabriel.Kaufmann@ericom.com> Ericom Software Tel (Dir): +972 2 591 1700 Ext 754 Tel (Main): +972 2 591 1700 http://www.ericom.com Access Done Right Empower Enterprise-Wide Access to Microsoft(r) Terminal Server, Virtual Desktops and Legacy Applications

2 1

[libvir] Could not connect to xend
by Anton Protopopov 11 Mar '08

11 Mar '08

Hi. When I tried to run virsh to connect to xen, I have got an error --- libvirt failed to open connection to xend. Here is output with DEBUG enabled. What kind of additional information do you need? [root@localhost ~]# LIBVIRT_DEBUG=SPAM virsh DEBUG: libvirt.c: virInitialize (register drivers) DEBUG: libvirt.c: virConnectOpenAuth (name=(null), auth=0xb7ed79d8, flags=0) DEBUG: libvirt.c: do_open (Probed xen:///) DEBUG: libvirt.c: do_open (Using xen:/// as default URI, 1 hypervisor found) DEBUG: libvirt.c: do_open (name "xen:///" to URI components: scheme xen opaque (null) authority (null) server (null) user (null) port 0 path / ) DEBUG: libvirt.c: do_open (trying driver 0 (Test) ...) DEBUG: libvirt.c: do_open (driver 0 Test returned DECLINED) DEBUG: libvirt.c: do_open (trying driver 1 (QEMU) ...) DEBUG: libvirt.c: do_open (driver 1 QEMU returned DECLINED) DEBUG: libvirt.c: do_open (trying driver 2 (Xen) ...) DEBUG: xen_unified.c: xenUnifiedOpen (trying Xen sub-driver 0) DEBUG: xen_unified.c: xenUnifiedOpen (Xen sub-driver 0 open ok ) DEBUG: xen_unified.c: xenUnifiedOpen (trying Xen sub-driver 2) libvir: Xen Daemon error : internal error failed to connect to xend libvir: Xen Daemon error : internal error failed to connect to xend DEBUG: xen_unified.c: xenUnifiedOpen (Xen sub-driver 2 open failed ) DEBUG: libvirt.c: do_open (driver 2 Xen returned ERROR) DEBUG: hash.c: virUnrefConnect (unref connection 0x807bf88 xen:/// 1) DEBUG: hash.c: virReleaseConnect (release connection 0x807bf88 xen:///) error: failed to connect to the hypervisor

2 1

[Libvir] Need Help on Error handling.
by Ajishrao.r 11 Mar '08

11 Mar '08

Hi, I am using libvirt to manage, The Hypervisor, Can Any one Help me about handling error? My main questions are. --- What is Library level Error Handling? --- Any Example to use void virSetErrorFunc (void * userData,virErrorFunc <http://libvirt.org/html/libvirt-virterror.html#virErrorFunc> handler) --- What is the difference between these two funtions? int virConnCopyLastError (virConnectPtr <http://libvirt.org/html/libvirt-libvirt.html#virConnectPtr> conn,virErrorPtr <http://libvirt.org/html/libvirt-virterror.html#virErrorPtr> to) virErrorPtr <http://libvirt.org/html/libvirt-virterror.html#virErrorPtr> virConnGetLastError (virConnectPtr <http://libvirt.org/html/libvirt-libvirt.html#virConnectPtr> conn) -- Thanks&Regards Ajeesh

3 3

[Libvir] PATCH: Don't make connection read-only when non-root
by Daniel P. Berrange 11 Mar '08

11 Mar '08

The virsh commands has long forced the connection to be read-only if running as non-root. This is bogus because it is perfectly capable of authenticating full read-write connections as non-root since we gained kerberos/policykit support. The user can always use the explicit --readonly flag if they only want a read only connection Dan. Index: src/virsh.c =================================================================== RCS file: /data/cvs/libvirt/src/virsh.c,v retrieving revision 1.135 diff -r1.135 virsh.c 6048,6054d6047 < #ifndef __MINGW32__ < /* Force a non-root, Xen connection to readonly */ < if ((ctl->name == NULL || < !strcasecmp(ctl->name, "xen")) && ctl->uid != 0) < ctl->readonly = 1; < #endif < -- |: Red Hat, Engineering, Boston -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 :|

3 2

[Libvir] PATCH: Fix libvirtd to not shutdown at startup
by Daniel P. Berrange 11 Mar '08

11 Mar '08

Changes to the daemon to make it called qemudShutdown() broke the network driver when run in --daemon mode. This is because the main daemon forks into the background, and the original PID then exits destroying any active networks & removing the iptables rules. # /usr/sbin/libvirtd --daemon Shutting down network 'default'iptables: No chain/target/match by that name iptables: Bad rule (does a matching rule exist in that chain?) iptables: Bad rule (does a matching rule exist in that chain?) iptables: Bad rule (does a matching rule exist in that chain?) iptables: No chain/target/match by that name iptables: No chain/target/match by that name iptables: Bad rule (does a matching rule exist in that chain?) iptables: Bad rule (does a matching rule exist in that chain?) iptables: Bad rule (does a matching rule exist in that chain?) iptables: Bad rule (does a matching rule exist in that chain?) Failed to bring down bridge 'virbr0' : No such device Failed to delete bridge 'virbr0' : No such device or address The core issue is that when forking into daemon mode the original process should immediately exit & not do invoke cleanup functions. This patch makes it do that Dan. Index: qemud/qemud.c =================================================================== RCS file: /data/cvs/libvirt/qemud/qemud.c,v retrieving revision 1.89 diff -u -r1.89 qemud.c --- qemud/qemud.c 3 Mar 2008 18:10:19 -0000 1.89 +++ qemud/qemud.c 9 Mar 2008 20:56:27 -0000 @@ -393,7 +393,7 @@ case -1: return -1; default: - return nextpid; + _exit(0); } cleanup: @@ -418,8 +418,7 @@ status != 0) { return -1; } - - return pid; + _exit(0); } } } @@ -2116,16 +2115,12 @@ goto error1; if (godaemon) { - int pid; openlog("libvirtd", 0, 0); - pid = qemudGoDaemon(); - if (pid < 0) { + if (qemudGoDaemon() < 0) { qemudLog(QEMUD_ERR, _("Failed to fork as daemon: %s"), strerror(errno)); goto error1; } - if (pid > 0) - goto out; /* Choose the name of the PID file. */ if (!pid_file) { @@ -2172,7 +2167,6 @@ if (godaemon) closelog(); - out: ret = 0; error2: -- |: Red Hat, Engineering, Boston -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 :|

3 2

Re: [Libvir] virsh start problem + patch
by Richard W.M. Jones 11 Mar '08

11 Mar '08

OK, I think that's a good explanation but I'll leave it to Dan to comment 'coz he wrote the code ... Rich. -- Richard Jones, Emerging Technologies, Red Hat http://et.redhat.com/~rjones virt-p2v converts physical machines to virtual machines. Boot with a live CD or over the network (PXE) and turn machines into Xen guests. http://et.redhat.com/~rjones/virt-p2v

1 0