October 2008 - Devel - libvirt List Archives

[libvirt] Xen and out of memory stuff, blocks libvirtd
by Stefan de Konink 09 Oct '08

09 Oct '08

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512 [2008-10-08 19:14:05 13001] ERROR (XendCheckpoint:157) Save failed on domain klant1_monetdb (17) - resuming. Traceback (most recent call last): File "usr/lib64/python2.5/site-packages/xen/xend/XendCheckpoint.py", line 125, in save forkHelper(cmd, fd, saveInputHandler, False) File "usr/lib64/python2.5/site-packages/xen/xend/XendCheckpoint.py", line 358, in forkHelper child = xPopen3(cmd, True, -1, [fd, xc.handle()]) File "usr/lib64/python2.5/site-packages/xen/util/xpopen.py", line 100, in __init__ self.pid = os.fork() OSError: [Errno 12] Cannot allocate memory [2008-10-08 19:14:05 13001] DEBUG (XendDomainInfo:2456) XendDomainInfo.resumeDomain(17) After this point, I am unable to contact the libvird daemon. Stefan -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEAREKAAYFAkjs7qQACgkQYH1+F2Rqwn0heQCeNJWDo2h3BR1IhcXmt+w9tpjD 304An1SlqWbig7nL8jgeElzJgu/KqgIa =dl3M -----END PGP SIGNATURE-----

2 7

[libvirt] [PATCH] Unify *ReportError logic
by Cole Robinson 09 Oct '08

09 Oct '08

Currently, most src/* files have their own ReportError function. Some support printf style arguments, others only allow reporting a single string message. The code for all of them does virtually the same thing, possibly passing a different constant off to another function. The attached patch adds a function to virterror.c which encapsulates the common ReportError logic. I used this to replace qemudReportError with a macro, which also allows passing off filename and line number info if we wanted to do something with it later. I did just the one function conversion to see what people think: if I'm missing something, or ideas for anything else to add. Seems to work as expected in my testing. Thanks, Cole

3 5

[libvirt] first argument in virDomainMigrate, is it free'ed?
by Stefan de Konink 09 Oct '08

09 Oct '08

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512 Hi, I saw that freeing the dom after a migration takes place results in: libvir: Domain error : invalid domain pointer in virDomainFree Now I wonder, if this is the case, could a normale migrate function 'just' be: domain = virDomainMigrate(domain, ....) Or is this unsafe; for example as a domain migration fails? Stefan -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEAREKAAYFAkjqOYwACgkQYH1+F2Rqwn0NZACffFESznbeElKNVgt65YXPGzXX ausAmwZ7HScE0/RTtDWm2uug/dC1S21K =xpn6 -----END PGP SIGNATURE-----

3 3

[libvirt] Xen Migration does not work
by Stefan de Konink 09 Oct '08

09 Oct '08

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512 I'm try to migrate using libvirt: virsh # migrate 0802367323_PIQ002 xen://xen002.local/ libvir: Xen Daemon error : POST operation failed: xend_post: error from xen daemon: (xend.err 'Missing parameter: node') virsh -v 0.4.6 Stefan -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEAREKAAYFAkjpagoACgkQYH1+F2Rqwn0yogCfd0+rK0mNrLpmmwclAjRpevlg w4EAn2VOWooBbch7awWpRJs/1Sn7G5ZW =6eZd -----END PGP SIGNATURE-----

3 5

[libvirt] [PATCH] Add generic parameter=value support for virsh's schedinfo command
by Dan Smith 09 Oct '08

09 Oct '08

This patch maintains the two Xen-specific --weight and --cap options, but adds support for setting arbitrary parameters by specifying them in param=value syntax. Changes to the virsh manual are included. Changes: - Replace use of 'a' conversion modifier with pre-alloc diff -r 51fd150edd42 -r e441b1a06f1c docs/virsh.pod --- a/docs/virsh.pod Tue Oct 07 11:45:52 2008 -0700 +++ b/docs/virsh.pod Wed Oct 08 07:59:11 2008 -0700 @@ -318,12 +318,14 @@ with all the same limitations. Open network connections may be severed upon restore, as TCP timeouts may have expired. +=item B<schedinfo> optional I<--set> B<parameter=value> I<domain-id> + =item B<schedinfo> optional I<--weight> B<number> optional I<--cap> B<number> I<domain-id> -Allows to show (and set) the domain scheduler parameters. This is currently -only defined for XEN_CREDIT scheduler, and the optional weight and cap -arguments allows to set the associated parameters in that scheduler if -provided. +Allows to show (and set) the domain scheduler parameters. + +B<Note>: The weight and cap parameters are defined only for the +XEN_CREDIT scheduler and are now I<DEPRECATED>. =item B<setmem> I<domain-id> B<kilobytes> diff -r 51fd150edd42 -r e441b1a06f1c src/virsh.c --- a/src/virsh.c Tue Oct 07 11:45:52 2008 -0700 +++ b/src/virsh.c Wed Oct 08 07:59:11 2008 -0700 @@ -1112,6 +1112,7 @@ static const vshCmdOptDef opts_schedinfo[] = { {"domain", VSH_OT_DATA, VSH_OFLAG_REQ, gettext_noop("domain name, id or uuid")}, + {"set", VSH_OT_STRING, VSH_OFLAG_NONE, gettext_noop("parameter=value")}, {"weight", VSH_OT_INT, VSH_OFLAG_NONE, gettext_noop("weight for XEN_CREDIT")}, {"cap", VSH_OT_INT, VSH_OFLAG_NONE, gettext_noop("cap for XEN_CREDIT")}, {NULL, 0, 0, NULL} @@ -1121,6 +1122,9 @@ cmdSchedinfo(vshControl *ctl, const vshCmd *cmd) { char *schedulertype; + char *set; + char *param_name = NULL; + long long int param_value = 0; virDomainPtr dom; virSchedParameterPtr params = NULL; int i, ret; @@ -1128,6 +1132,7 @@ int nr_inputparams = 0; int inputparams = 0; int weightfound = 0; + int setfound = 0; int weight = 0; int capfound = 0; int cap = 0; @@ -1141,7 +1146,7 @@ if (!(dom = vshCommandOptDomain(ctl, cmd, "domain", NULL))) return FALSE; - /* Currently supports Xen Credit only */ + /* Deprecated Xen-only options */ if(vshCommandOptBool(cmd, "weight")) { weight = vshCommandOptInt(cmd, "weight", &weightfound); if (!weightfound) { @@ -1160,6 +1165,25 @@ } else { nr_inputparams++; } + } + + if(vshCommandOptBool(cmd, "set")) { + set = vshCommandOptString(cmd, "set", &setfound); + if (!setfound) { + vshError(ctl, FALSE, "%s", _("Error getting param")); + goto cleanup; + } + + param_name = vshMalloc(ctl, strlen(set) + 1); + if (param_name == NULL) + goto cleanup; + + if (sscanf(set, "%[^=]=%i", param_name, &param_value) != 2) { + vshError(ctl, FALSE, "%s", _("Invalid value of param")); + goto cleanup; + } + + nr_inputparams++; } params = vshMalloc(ctl, sizeof (virSchedParameter) * nr_inputparams); @@ -1180,7 +1204,14 @@ params[inputparams].value.ui = cap; inputparams++; } - /* End Currently supports Xen Credit only */ + /* End Deprecated Xen-only options */ + + if (setfound) { + strncpy(params[inputparams].field,param_name,sizeof(params[0].field)); + params[inputparams].type = VIR_DOMAIN_SCHED_FIELD_LLONG; + params[inputparams].value.l = param_value; + inputparams++; + } assert (inputparams == nr_inputparams); @@ -1247,6 +1278,7 @@ } cleanup: free(params); + free(param_name); virDomainFree(dom); return ret_val; }

2 1

[libvirt] [PATCH] Add generic parameter=value support for virsh's schedinfo command
by Dan Smith 08 Oct '08

08 Oct '08

This patch maintains the two Xen-specific --weight and --cap options, but adds support for setting arbitrary parameters by specifying them in param=value syntax. Changes to the virsh manual are included. diff -r 51fd150edd42 -r aae58f42bd4a docs/virsh.pod --- a/docs/virsh.pod Tue Oct 07 11:45:52 2008 -0700 +++ b/docs/virsh.pod Tue Oct 07 13:53:49 2008 -0700 @@ -318,12 +318,14 @@ with all the same limitations. Open network connections may be severed upon restore, as TCP timeouts may have expired. +=item B<schedinfo> optional I<--set> B<parameter=value> I<domain-id> + =item B<schedinfo> optional I<--weight> B<number> optional I<--cap> B<number> I<domain-id> -Allows to show (and set) the domain scheduler parameters. This is currently -only defined for XEN_CREDIT scheduler, and the optional weight and cap -arguments allows to set the associated parameters in that scheduler if -provided. +Allows to show (and set) the domain scheduler parameters. + +B<Note>: The weight and cap parameters are defined only for the +XEN_CREDIT scheduler and are now I<DEPRECATED>. =item B<setmem> I<domain-id> B<kilobytes> diff -r 51fd150edd42 -r aae58f42bd4a src/virsh.c --- a/src/virsh.c Tue Oct 07 11:45:52 2008 -0700 +++ b/src/virsh.c Tue Oct 07 13:53:49 2008 -0700 @@ -1112,6 +1112,7 @@ static const vshCmdOptDef opts_schedinfo[] = { {"domain", VSH_OT_DATA, VSH_OFLAG_REQ, gettext_noop("domain name, id or uuid")}, + {"set", VSH_OT_STRING, VSH_OFLAG_NONE, gettext_noop("parameter=value")}, {"weight", VSH_OT_INT, VSH_OFLAG_NONE, gettext_noop("weight for XEN_CREDIT")}, {"cap", VSH_OT_INT, VSH_OFLAG_NONE, gettext_noop("cap for XEN_CREDIT")}, {NULL, 0, 0, NULL} @@ -1121,6 +1122,9 @@ cmdSchedinfo(vshControl *ctl, const vshCmd *cmd) { char *schedulertype; + char *set; + char *param_name = NULL; + long long int param_value = 0; virDomainPtr dom; virSchedParameterPtr params = NULL; int i, ret; @@ -1128,6 +1132,7 @@ int nr_inputparams = 0; int inputparams = 0; int weightfound = 0; + int setfound = 0; int weight = 0; int capfound = 0; int cap = 0; @@ -1141,7 +1146,7 @@ if (!(dom = vshCommandOptDomain(ctl, cmd, "domain", NULL))) return FALSE; - /* Currently supports Xen Credit only */ + /* Deprecated Xen-only options */ if(vshCommandOptBool(cmd, "weight")) { weight = vshCommandOptInt(cmd, "weight", &weightfound); if (!weightfound) { @@ -1160,6 +1165,21 @@ } else { nr_inputparams++; } + } + + if(vshCommandOptBool(cmd, "set")) { + set = vshCommandOptString(cmd, "set", &setfound); + if (!setfound) { + vshError(ctl, FALSE, "%s", _("Error getting param")); + goto cleanup; + } + + if (sscanf(set, "%a[^=]=%i", &param_name, &param_value) != 2) { + vshError(ctl, FALSE, "%s", _("Invalid value of param")); + goto cleanup; + } + + nr_inputparams++; } params = vshMalloc(ctl, sizeof (virSchedParameter) * nr_inputparams); @@ -1180,7 +1200,14 @@ params[inputparams].value.ui = cap; inputparams++; } - /* End Currently supports Xen Credit only */ + /* End Deprecated Xen-only options */ + + if (setfound) { + strncpy(params[inputparams].field,param_name,sizeof(params[0].field)); + params[inputparams].type = VIR_DOMAIN_SCHED_FIELD_LLONG; + params[inputparams].value.l = param_value; + inputparams++; + } assert (inputparams == nr_inputparams); @@ -1247,6 +1274,7 @@ } cleanup: free(params); + free(param_name); virDomainFree(dom); return ret_val; }

2 2

[libvirt] [PATCH] Fix qemudDomainShutdown() memory leak
by Eduardo Habkost 08 Oct '08

08 Oct '08

qemudDomainShutdown() currently doesn't free the returned reply from qemudMonitorCommand(). Signed-off-by: Eduardo Habkost <ehabkost(a)redhat.com> --- src/qemu_driver.c | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/src/qemu_driver.c b/src/qemu_driver.c index 806608d..a88cb75 100644 --- a/src/qemu_driver.c +++ b/src/qemu_driver.c @@ -2176,6 +2176,7 @@ static int qemudDomainShutdown(virDomainPtr dom) { "%s", _("shutdown operation failed")); return -1; } + VIR_FREE(info); return 0; } -- 1.5.5.GIT

2 1

[libvirt] 141 format string warnings
by Jim Meyering 08 Oct '08

08 Oct '08

Hello, I count 141 warnings like this: util.c:177: warning: format not a string literal and no format arguments If anyone is in the process of fixing those, please let me know soon.

3 3

[libvirt] [PATCH] (virCgroupSetValueStr): Use safe_write, not write.
by Jim Meyering 08 Oct '08

08 Oct '08

"make syntax-check" was failing. This fixes it: >From 155aea5141553fd9fab19aedc4df1613abae771c Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Tue, 7 Oct 2008 23:10:03 +0200 Subject: [PATCH] Tue Oct 7 23:08:51 CEST 2008 Jim Meyering <meyering(a)redhat.com> * src/cgroup.c (virCgroupSetValueStr): Use safe_write, not write. --- ChangeLog | 4 ++++ src/cgroup.c | 2 +- 2 files changed, 5 insertions(+), 1 deletions(-) diff --git a/ChangeLog b/ChangeLog index 619d2b6..88b746e 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,3 +1,7 @@ +Tue Oct 7 23:08:51 CEST 2008 Jim Meyering <meyering(a)redhat.com> + + * src/cgroup.c (virCgroupSetValueStr): Use safe_write, not write. + Tue Oct 7 18:33:39 CEST 2008 Daniel Veillard <veillard(a)redhat.com> * src/qemu_driver.c: another OOM handling cleanup (Guido Günther) diff --git a/src/cgroup.c b/src/cgroup.c index a7d2ca9..ed206dc 100644 --- a/src/cgroup.c +++ b/src/cgroup.c @@ -649,7 +649,7 @@ int virCgroupAddTask(virCgroupPtr group, pid_t pid) goto done; } - if (write(fd, pidstr, strlen(pidstr)) <= 0) { + if (safe_write(fd, pidstr, strlen(pidstr)) <= 0) { rc = -errno; goto done; } -- 1.6.0.2.307.gc4275

3 2

[libvirt] virsh connect
by devi thapa 08 Oct '08

08 Oct '08

Hi All, I have created a virtual network using virsh. I want to connect to the virtual machine, without being asked password. Can I achieve this using command virsh connect <name> Please help me out with this. Regards, Devi.

2 1

[libvirt] [PATCH] Reboot support for QEMU/KVM
by Eduardo Habkost 08 Oct '08

08 Oct '08

This patch implements reboot support on qemu_driver. It is so simple that I am wondering if there is some problem I am missing, that would be the reason it wasn't implemented yet. I hope not. :) Signed-off-by: Eduardo Habkost <ehabkost(a)redhat.com> --- src/qemu_driver.c | 23 ++++++++++++++++++++++- 1 files changed, 22 insertions(+), 1 deletions(-) diff --git a/src/qemu_driver.c b/src/qemu_driver.c index a88cb75..9ce53ad 100644 --- a/src/qemu_driver.c +++ b/src/qemu_driver.c @@ -2182,6 +2182,27 @@ static int qemudDomainShutdown(virDomainPtr dom) { } +static int qemudDomainReboot(virDomainPtr dom) { + struct qemud_driver *driver = (struct qemud_driver *)dom->conn->privateData; + virDomainObjPtr vm = virDomainFindByID(driver->domains, dom->id); + char* info; + + if (!vm) { + qemudReportError(dom->conn, dom, NULL, VIR_ERR_INVALID_DOMAIN, + _("no domain with matching id %d"), dom->id); + return -1; + } + + if (qemudMonitorCommand(driver, vm, "system_reset", &info) < 0) { + qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED, + "%s", _("reboot operation failed")); + return -1; + } + VIR_FREE(info); + return 0; +} + + static int qemudDomainDestroy(virDomainPtr dom) { struct qemud_driver *driver = (struct qemud_driver *)dom->conn->privateData; virDomainObjPtr vm = virDomainFindByID(driver->domains, dom->id); @@ -4032,7 +4053,7 @@ static virDriver qemuDriver = { qemudDomainSuspend, /* domainSuspend */ qemudDomainResume, /* domainResume */ qemudDomainShutdown, /* domainShutdown */ - NULL, /* domainReboot */ + qemudDomainReboot, /* domainReboot */ qemudDomainDestroy, /* domainDestroy */ qemudDomainGetOSType, /* domainGetOSType */ qemudDomainGetMaxMemory, /* domainGetMaxMemory */ -- 1.5.5.GIT

2 2

[libvirt] [PATCH 0 of 3] Add cpu.shares support to LXC driver
by Dan Smith 08 Oct '08

08 Oct '08

This set adds {Get,Set}SchedulerParameters support to the LXC driver. A single parameter of "cpu_shares" is added, which gets/sets the cpu.shares key in the domain's cgroup. I also added bits to "virsh schedinfo" to support this.

2 7

[libvirt] [PATCH]: (trivial) handle OOM as everywhere else
by Guido Günther 08 Oct '08

08 Oct '08

Hi, we print an error message on OOM conditions in qemudStartup() at the end of the function, no need to do it differently in one place. Cheers, -- Guido

2 1

[libvirt] [PATCH]: (trivial) getpwuid is not OOM
by Guido Günther 08 Oct '08

08 Oct '08

Hi, we shouldn't print an oom error message when actually getpwuid failed. Cheers, -- Guido

2 1

[libvirt] When bridge doesn't exist
by Anton Protopopov 07 Oct '08

07 Oct '08

Assume that we want to add a bridge to LAN to a guest. We saying to libvirt: ... <interface type="bridge"> <source bridge="BRIDGE"> </interface> ... What is the right behaviour of libvirt in case bridge BRIDGE doesn't exist---return an error or try to create that bridge? If it is the former, then who must create that bridge?

1 0

[libvirt] Routed Networking - Ubuntu Hardy KVM and LibVirt
by Barry Robinson 07 Oct '08

07 Oct '08

Gday, I have been searching for information on how to configure KVM guest OS' to do simple routing from their associated vnetX address to the host machine's ethernet network subnet (lets say eth0/192.168.0.0/24). I understand that by default the clients are NAT'ed to the Host's adapter. This is a handy feature, but it does not suit the setup I am trying to apply. Basically, I am trying to do simple routing from the host subnet (192.168.0.0/24) to the client's subnet (192.168.17.0/24), and back again. So I can talk to 192.168.0.0/24 addresses and they can talk to 192.168.17.0/24 (routing rule has been added so this is possible). There should be no NAT involved. The following is about as close as I have come to getting this solution. However, I still need to add in the iptables -A FORWARD -i vnetX -o eth0 -j ACCEPT rule in after the machine is booted. Is it possible to do this as part of the script process, or is there a post-ifup script that can be run? /etc/libvirt/qemu/network/routed-net.xml: <network> <name>routed-net</name> <uuid></uuid> <bridge name="virbr%d" /> <ip address="192.168.32.1" netmask="255.255.255.0"> <dhcp> <range start="192.168.32.33" end="192.168.32.63" /> </dhcp> </ip> </network> I have tried using the <forward='routed' /> option (under the bridge name option). This does add the forward rules, but for some reason still applies the MASQUERADE rule to the 192.168.17.0/24 network. (eg. 0 0 MASQUERADE all -- * * 192.168.32.0/24 0.0.0.0/0 ) libvirtd seems to hide it's firewall rulesets pretty well, cause I can't even find them to manually add/remove rules. Not in any place obvious (that I can find), like /etc. Any help would be appreciated. Apologies if this topic has been covered... can't find it anywhere using Google. Cheers, Barry Linseed Technologies Open Source IT Solutions Phone: 0415131452 Email: brobinson(a)linseed.com.au Website: http://www.linseed.com.au

2 1

[libvirt] PATCH: 0 of 5: Remove all linked lists
by Daniel P. Berrange 07 Oct '08

07 Oct '08

I've been doing some proof of concept work to make the libvirtd daemon multi-threaded, and this in turns mean that the QEMU / LXC / OpenVZ drivers need to have some degree of locking on their internal data structures. Unforatunately the internal domain/network/storage APIs make extensive use of linked lists for tracking objects. This makes it incredibly hard to do fine grained locking of individual objects - eg, deleting one object requires locking the object it points to, and one the pointing to it in the linked list, as well as the list itself. This would be so complex we'd inevitably introduce locking bugs. So the series of patches which follow replace all linked lists with explicitly sized arrays. This means I only need 2 levels of locks, one on the driver object, and one of the domain/network/storage object being manipulated. The semantics for this are nice & easy to define. Daniel -- |: Red Hat, Engineering, London -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 :|

2 13

[libvirt] [PATCH] Centralize use of DEBUG macros
by Cole Robinson 07 Oct '08

07 Oct '08

Currently the DEBUG and DEBUG0 macros are duplicated in every file that uses them. This patch moves the macros to internal.h, removes the needless duplication, and now every file gets them for free. Seems to work as expected in my testing. Thanks, Cole

3 6

[libvirt] [PATCH] Slim xmllint test output
by Cole Robinson 07 Oct '08

07 Oct '08

This patch simply slims down the output from the xmllint tests to not overrun a single line, only printing the relevant information about the xml file (parent directory and name, not fully qualified path). Purely cosmetic, but more useful IMO. Thanks, Cole

2 2

[libvirt] [PATCH] Fix possible segfault if starting qemu vm with an inactive virtual network
by Cole Robinson 07 Oct '08

07 Oct '08

The attached patch fixes a possible segfault if trying to start a domain which tries to use an inactive virtual network. We should explicitly catch that error and report it, but that's a separate patch. Thanks, Cole

3 3

[libvirt] [PATCH] use C99 initializers for virState
by Guido Günther 07 Oct '08

07 Oct '08

Hi, attached patch uses C99 initializers for virStateDriver. Makes reading easier and extending the structure less error prone. -- Guido

3 2

[libvirt] [PATCH] fix build in separate build-dir
by Guido Günther 06 Oct '08

06 Oct '08

Hi, building in a separate build dir is currently broken since the Makefile is looking for libvirt.policy in the wrong location. Fix attached. Cheers, -- Guido

2 1

[libvirt] [discuss] The new cgroup patches for libvirt
by Balbir Singh 04 Oct '08

04 Oct '08

Hi, Everyone, I've seen a new set of patches from Dan Smith, which implement cgroup support for libvirt. While the patches seem simple, there are some issues that have been pointed out in the posting itself. I hope that libvirt will switch over (may be after your concerns are addressed and definitely in the longer run) to using libcgroups rather than having an internal implementation of cgroups. The advantages of switching over would be using the functionality that libcgroup already provides libcgroups (libcg.sf.net) provides 1. Ability to configure and mount cgroups and controllers via initscripts and a configuration file 2. An API to control and read cgroups information 3. Thread safety around API calls 4. Daemons to automatically classify a task based on a certain set of rules 5. API to extract current cgroup classification (where is the task currently in the cgroup hierarchy) While re-implementing might sound like a cool thing to do, here are the drawbacks 1. It leads to code duplication and reduces code reuse 2. It leads to confused users I understand that in the past there has been a perception that libcgroups might not yet be ready, because we did not have ABI stability built into the library and the header file had old comments about things changing. I would urge the group to look at the current implementation of libcgroups (look at v0.32) and help us 1. Fix any issues you see or point them to us 2. Add new API or request for new API that can help us integrate better with libvirt -- Balbir

4 11

[libvirt] [PATCH 0 of 2] [RFC] Add cgroup manipulation and LXC driver support
by Dan Smith 04 Oct '08

04 Oct '08

This patch set adds basic cgroup support to the LXC driver. It consists of a small internal cgroup manipulation API, as well as changes to the driver itself to utilize the support. Currently, we just set a memory limit and the allowed devices list. The cgroup.{c,h} interface can be easily redirected to libcgroup in the future if and when the decision to move in that direction is made. Some discussion on the following points is probably warranted, to help determine how deep we want to go with this internal implementation, in terms' of supporting complex system configurations, etc. - What to do if controllers are mounted in multiple places - What to do if memory and device controllers aren't present - What to do if the root group is set for exclusive cpuset behavior Some additional caveats are noted per-patch as well.

3 16

[libvirt] [PATCH 0 of 2] Add cgroup manipulation and LXC driver support
by Dan Smith 04 Oct '08

04 Oct '08

This is my updated set for cgroup manipulation. I think I have addressed all of the concerns and comments so far. There are no major caveats I'm aware of with this round.

3 15