There is a race condition in the monitoring of the migration process on KVM. If the monitor wakes up in the tight window after the migration succeeds, but before the migration thread terminates, the monitor will get a LibvirtException “Domain not found: no domain with matching uuid” when checking on the migration status. This in turn causes CloudStack to sync the VM state to stop, in which it issues a defensive StopCommand to ensure it is correctly synced.
Fix: Prevent LibvirtException: "Domain not found" caused by the call to dm.getInfo()
The documentation of Libvirt specifies the requirement of using an XML namespace,
when having metadata in the Domain XML. The Nuage extenstion metadata was not
adhering to this specification, and the lastest Libvirt version ignores it in that case.
Allowed zone-wide primary storage based on a custom plug-in to be added via the GUI in a KVM-only environment (previously this only worked for XenServer and VMware)
Added support for root disks on managed storage with KVM
Added support for volume snapshots with managed storage on KVM
Enable creating a template directly from a volume (i.e. without having to go through a volume snapshot) on KVM with managed storage
Only allow the resizing of a volume for managed storage on KVM if the volume in question is either not attached to a VM or is attached to a VM in the Stopped state.
Included support for Reinstall VM on KVM with managed storage
Enabled offline migration on KVM from non-managed storage to managed storage and vice versa
Included support for online storage migration on KVM with managed storage (NFS and Ceph to managed storage)
Added support to download (extract) a managed-storage volume to a QCOW2 file
When uploading a file from outside of CloudStack to CloudStack, set the min and max IOPS, if applicable.
Included support for the KVM auto-convergence feature
The compression flag was actually added in version 1.0.3 (1000003) as opposed to version 1.3.0 (1003000) (changed this to reflect the correct version)
On KVM when using iSCSI-based managed storage, if the user shuts a VM down from the guest OS (as opposed to doing so from CloudStack), we need to pass to the KVM agent a list of applicable iSCSI volumes that need to be disconnected.
Added a new Global Setting: kvm.storage.live.migration.wait
For XenServer, added a check to enforce that only volumes from zone-wide managed storage can be storage motioned from a host in one cluster to a host in another cluster (cannot do so at the time being with volumes from cluster-scoped managed storage)
Don’t allow Storage XenMotion on a VM that has any managed-storage volume with one or more snapshots.
Enabled for managed storage with VMware: Template caching, create snapshot, delete snapshot, create volume from snapshot, and create template from snapshot
Added an SIOC API plug-in to support VMware SIOC
When starting a VM that uses managed storage in a cluster other than the one it last was running in, we need to remove the reference to the iSCSI volume from the original cluster.
Added the ability to revert a volume to a snapshot
Enabled cluster-scoped managed storage
Added support for VMware dynamic discovery
Extending Config Drive support
* Added support for VMware
* Build configdrive.iso on ssvm
* Added support for VPC and Isolated Networks
* Moved implementation to new Service Provider
* UI fix: add support for urlencoded userdata
* Add support for building systemvm behind a proxy
Co-Authored-By: Raf Smeets <raf.smeets@nuagenetworks.net>
Co-Authored-By: Frank Maximus <frank.maximus@nuagenetworks.net>
Co-Authored-By: Sigert Goeminne <sigert.goeminne@nuagenetworks.net>
This feature allows using templates and ISOs avoiding secondary storage as intermediate cache on KVM. The virtual machine deployment process is enhanced to supported bypassed registered templates and ISOs, delegating the work of downloading them to primary storage to the KVM agent instead of the SSVM agent.
Template and ISO registration:
- When hypervisor is KVM, a checkbox is displayed with 'Direct Download' label.
- API methods registerTemplate and registerISO are both extended with this new parameter directdownload.
- On template or ISO registration, no download job is sent to SSVM agent, CloudStack would only persist an entry on template_store_ref indicating that template or ISO has been marked as 'Direct Download' (bypassing Secondary Storage). These entries are persisted as:
template_id = Template or ISO id on vm_template table
store_id NULL
download_state = BYPASSED
state = Ready
(Note: these entries allow users to deploy virtual machine from registered templates or ISOs)
- An URL validation command is sent to a random KVM host to check if template/ISO location can be reached. Metalink are also supported by this feature. In case of a metalink, it is fetched and URL check is performed on each of its URLs.
- Checksum should be provided as indicated on #2246: {ALGORITHM}CHKSUMHASH
- After template or ISO is registered, it would be displayed in the UI
Virtual machine deployment:
When a 'Direct Download' template is selected for deployment, CloudStack would delegate template downloading to destination storage pool via destination host by a new pluggable download manager.
Download manager would handle template downloading depending on URL protocol. In case of HTTP, request headers can be set by the user via vm_template_details. Those details should be persisted as:
Key: HTTP_HEADER
Value: HEADERNAME:HEADERVALUE
In case of HTTPS, a new API method is added uploadTemplateDirectDownloadCertificate to allow user importing a client certificate into all KVM hosts' keystore before deployment.
After template or ISO is downloaded to primary storage, usual entry would be persisted on template_spool_ref indicating the mapping between template/ISO and storage pool.
* Cleanup and Improve NetUtils
This class had many unused methods, inconsistent names and redundant code.
This commit cleans up code, renames a few methods and constants.
The global/account setting 'api.allowed.source.cidr.list' is set
to 0.0.0.0/0,::/0 by default preserve the current behavior and thus
allow API calls for accounts from all IPv4 and IPv6 subnets.
Users can set it to a comma-separated list of IPv4/IPv6 subnets to
restrict API calls for Admin accounts to certain parts of their network(s).
This is to improve Security. Should an attacker steal the Access/Secret key
of an account he/she still needs to be in a subnet from where accounts are
allowed to perform API calls.
This is a good security measure for APIs which are connected to the public internet.
Signed-off-by: Wido den Hollander <wido@widodh.nl>
- Refactors and simplifies systemvm codebase file structures keeping
the same resultant systemvm.iso packaging
- Password server systemd script and new postinit script that runs
before sshd starts
- Fixes to keepalived and conntrackd config to make rVRs work again
- New /etc/issue featuring ascii based cloudmonkey logo/message and
systemvmtemplate version
- SystemVM python codebase linted and tested. Added pylint/pep to
Travis.
- iptables re-application fixes for non-VR systemvms.
- SystemVM template build fixes.
- Default secondary storage vm service offering boosted to have 2vCPUs
and RAM equal to console proxy.
- Fixes to several marvin based smoke tests, especially rVR related
tests. rVR tests to consider 3*advert_int+skew timeout before status
is checked.
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
This ports PR #1470 by @remibergsma.
Make the generated json files unique to prevent concurrency issues:
The json files now have UUIDs to prevent them from getting overwritten
before they've been executed. Prevents config to be pushed to the wrong
router.
2016-02-25 18:32:23,797 DEBUG [c.c.a.t.Request] (AgentManager-Handler-1:null) (logid:) Seq 2-4684025087442026584: Processing: { Ans: , MgmtId: 90520732674657, via: 2, Ver: v1, Flags: 10, [{"com.cloud.agent.api.routing.GroupA
nswer":{"results":["null - success: null","null - success: [INFO] update_config.py :: Processing incoming file => vm_dhcp_entry.json.4ea45061-2efb-4467-8eaa-db3d77fb0a7b\n[INFO] Processing JSON file vm_dhcp_entry.json.4ea4506
1-2efb-4467-8eaa-db3d77fb0a7b\n"],"result":true,"wait":0}}] }
On the router:
2016-02-25 18:32:23,416 merge.py __moveFile:298 Processed file written to /var/cache/cloud/processed/vm_dhcp_entry.json.4ea45061-2efb-4467-8eaa-db3d77fb0a7b.gz
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Otherwise we send down a 'null' to a ProcessBuilder in Java instead of a String and this
causes a NPE.
We should check first if the Instance has a IPv6 address before sending it there.
Signed-off-by: Wido den Hollander <wido@widodh.nl>
* CLOUDSTACK-10160: Fix typo in Libvirt XML definition for Virtio-SCSI
The attribute for the XML element 'controller' should be 'model' and
not 'mode'.
Source: https://libvirt.org/formatdomain.html#elementsControllers
A scsi controller has an optional attribute model, which is one of
'auto', 'buslogic', 'ibmvscsi', 'lsilogic', 'lsisas1068', 'lsisas1078',
'virtio-scsi' or 'vmpvscsi'.
In the current state a regular SCSI device is attached and not a Virtio-SCSI
device.
Signed-off-by: Wido den Hollander <wido@widodh.nl>
* CLOUDSTACK-10160: Add UnitTest for LibvirtVMDef.SCSIDef
To make sure the XML output string is correct
Signed-off-by: Wido den Hollander <wido@widodh.nl>
This commit adds support for passing IPv6 Addresses and/or Subnets as
Secondary IPs.
This is groundwork for CLOUDSTACK-9853 where IPv6 Subnets have to be
allowed in the Security Groups of Instances to we can add DHCPv6
Prefix Delegation.
Use ; instead of : for separating addresses, otherwise it would cause
problems with IPv6 Addresses.
Signed-off-by: Wido den Hollander <wido@widodh.nl>
* CLOUDSTACK-9972: Enhance listVolume API to include physical size and utilization.
Also fixed pool, cluster and pod info
* CLOUDSTACK-9972: Fix volume_view and duplicate API constant
* CLOUDSTACK-9972: Backport Do not allow vms to be deployed on hosts that are in disabled pod
* CLOUDSTACK-9972: Fix localization missing keys
* CLOUDSTACK-9972: Fix sql path
Commit enables a new feature for KVM hypervisor which purpose is to increase virtually amount of RAM available beyond the actual limit.
There is a new parameter in agent.properties: host.overcommit.mem.mb which enables adding specified amount of RAM to actually available. It is necessary to utilize KSM and ZSwap features which extend RAM with deduplication and compression.
The watchdog timer adds functionality where the Hypervisor can detect if an
instance has crashed or stopped functioning.
The watchdog timer adds functionality where the Hypervisor can detect if an
instance has crashed or stopped functioning.
When the Instance has the 'watchdog' daemon running it will send heartbeats
to the /dev/watchdog device.
If these heartbeats are no longer received by the HV it will reset the Instance.
If the Instance never sends the heartbeats the HV does not take action. It only
takes action if it stops sending heartbeats.
This is supported since Libvirt 0.7.3 and can be defined in the XML format as
described in the docs: https://libvirt.org/formatdomain.html#elementsWatchdog
To the 'devices' section this will be added:
In the agent.properties the action to be taken can be defined:
vm.watchdog.action=reset
The same goes for the model. The Intel i6300esb is however the most commonly used.
vm.watchdog.model=i6300esb
When the Instance has the 'watchdog' daemon running it will send heartbeats
to the /dev/watchdog device.
If these heartbeats are no longer received by the HV it will reset the Instance.
If the Instance never sends the heartbeats the HV does not take action. It only
takes action if it stops sending heartbeats.
This is supported since Libvirt 0.7.3 and can be defined in the XML format as
described in the docs: https://libvirt.org/formatdomain.html#elementsWatchdog
To the 'devices' section this will be added:
<watchdog model='i6300esb' action='reset'/>
In the agent.properties the action to be taken can be defined:
vm.watchdog.action=reset
The same goes for the model. The Intel i6300esb is however the most commonly used.
vm.watchdog.model=i6300esb
Signed-off-by: Wido den Hollander <wido@widodh.nl>
- Removed three bg thread tasks, uses FSM event-trigger based scheduling
- On successful recovery, kicks VM HA
- Improves overall HA scheduling and task submission, lower DB access
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Host-HA offers investigation, fencing and recovery mechanisms for host that for
any reason are malfunctioning. It uses Activity and Health checks to determine
current host state based on which it may degrade a host or try to recover it. On
failing to recover it, it may try to fence the host.
The core feature is implemented in a hypervisor agnostic way, with two separate
implementations of the driver/provider for Simulator and KVM hypervisors. The
framework also allows for implementation of other hypervisor specific provider
implementation in future.
The Host-HA provider implementation for KVM hypervisor uses the out-of-band
management sub-system to issue IPMI calls to reset (recover) or poweroff (fence)
a host.
The Host-HA provider implementation for Simulator provides a means of testing
and validating the core framework implementation.
Signed-off-by: Abhinandan Prateek <abhinandan.prateek@shapeblue.com>
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Since libvirt 1.2.2 libvirt will properly create volumes
using RBD format 2.
We can use libvirt to creates the volumes which strips a bit of
code from the CloudStack Agent's responsbility.
RBD format 2 is already used by all volumes created by CloudStack.
This format is the most recent format of RBD and is still actively
being developed.
This removes the support for Ubuntu 12.04 as that does not have the
proper libvirt version available.
Signed-off-by: Wido den Hollander wido@widodh.nl
We can use libvirt to creates the volumes which strips a bit of
code from the CloudStack Agent's responsbility.
RBD format 2 is already used by all volumes created by CloudStack.
This format is the most recent format of RBD and is still actively
being developed.
This removes the support for Ubuntu 12.04 as that does not have the
proper libvirt version available.
Signed-off-by: Wido den Hollander <wido@widodh.nl>
Updated hardcoded value with max data volumes limit from hypervisor capabilities.
(cherry picked from commit 93f5b6e8a391ce8b09be484d029c54d48a2b88aa)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Libvirt / Qemu (KVM) does not collect statistics about these either.
On some systems it might even yield a 'internal error' from libvirt
when attempting to gather block statistics from such devices.
For example Ubuntu 16.04 (Xenial) has a issue with this.
Skip them when looping through all devices.
Signed-off-by: Wido den Hollander <wido@widodh.nl>
The 'force' option provided with the stopVirtualMachine API command is
often assumed to be a hard shutdown sent to the hypervisor, when in fact
it is for CloudStacks' internal use. CloudStack should be able to send
the 'hard' power-off request to the hosts.
When forced parameter on the stopVM API is true, power off (hard shutdown)
a VM. This uses initial changes from #1635 to pass the forced parameter
to hypervisor plugin via the StopCommand, and fixes force stop (poweroff)
handling for KVM, VMware and XenServer.
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
