If the libvirt mount point is still busy and can't be unmounted
right now, it was waited 5 seconds and an plain unmount was tried,
without cleaning up the libvirt storagepool.
This kept libvirt thinking the storagepool
is active and mounted (which it wasn't).
Now after the plain unmount call, also
the libvirt storagepool will be destroyed.
- mTLS implementation for cluster service communication
- Listen only on the specified cluster node IP address instead of all interfaces
- Validate incoming cluster service requests are from peer management servers based on the server's certificate dns name which can be through global config - ca.framework.cert.management.custom.san
- Hardening of KVM command wrapper script exeicution
- Improve API server integration port check
- cloudstack-management.default: don't have JMX configuration if not needed. JMX is used for instrumentation; users who need to use it should enable it explicitly
Co-authored-by: Abhishek Kumar <abhishek.mrt22@gmail.com>
Co-authored-by: Wei Zhou <weizhou@apache.org>
Co-authored-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Signed-off-by: Abhishek Kumar <abhishek.mrt22@gmail.com>
- mTLS implementation for cluster service communication
- Listen only on the specified cluster node IP address instead of all interfaces
- Validate incoming cluster service requests are from peer management servers based on the server's certificate dns name which can be through global config - ca.framework.cert.management.custom.san
- Hardening of KVM command wrapper script execution
- Improve API server integration port check
- cloudstack-management.default: don't have JMX configuration if not needed. JMX is used for instrumentation; users who need to use it should enable it explicitly
Co-authored-by: Abhishek Kumar <abhishek.mrt22@gmail.com>
Co-authored-by: Wei Zhou <weizhou@apache.org>
Co-authored-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Signed-off-by: Abhishek Kumar <abhishek.mrt22@gmail.com>
* Mitigation for non-scalable Powerflex/ScaleIO clients
- Added ScaleIOSDCManager to manage SDC connections, checks clients limit, prepare and unprepare SDC on the hosts.
- Added commands for prepare and unprepare storage clients to prepare/start and stop SDC service respectively on the hosts.
- Introduced config 'storage.pool.connected.clients.limit' at storage level for client limits, currently support for Powerflex only.
* tests issue fixed
* refactor / improvements
* lock with powerflex systemid while checking connections limit
* updated powerflex systemid lock to hold till sdc preparation
* Added custom stats support for storage pool, through listStoragePools API
* code improvements, and unit tests
* unit tests fixes
* Update config 'storage.pool.connected.clients.limit' to dynamic, and some improvements
* Stop SDC on host after migration if no volumes mapped to host
* Wait for SDC to connect after scini service start, and some log improvements
* Do not throw exception (log it) when SDC is not connected while revoking access for the powerflex volume
* some log improvements
* Create/Export OVA file of the VM on external vCenter host, to temporary conversion location (NFS)
* Fixed ova issue on untar/extract ovf from ova file
"tar -xf" cmd on ova fails with "ovf: Not found in archive" while extracting ovf file
* Updated VMware to KVM instance migration using OVA
* Refactoring and cleanup
* test fixes
* Consider zone wide pools in the destination cluster for instance conversion
* Remove local storage pool support as temporary conversion location
- OVA export not possible as the pool is not accessible outside host, NFS pools are supported.
* cleanup unused code
* some improvements, and refactoring
* import nic unit tests
* vmware guru unit tests
* Separate clone VM and create template file for VMware migration
- Export OVA (of the cloned VM) to the conversion location takes time.
- Do any validations with cloned VM before creating the template (and fail early).
- Updated unit tests.
* Check conversion support on host before clone vm / create template on vmware (and fail early)
* minor code improvements
* Auto select the host with instance conversion capability
* Skip instance conversion supported response param for non-KVM hosts
* Show supported conversion hosts in the UI
* Skip persistence map update if network doesn't exist
* Added support to export OVA from KVM host, through ovftool (when installed in KVM host)
* Updated importvm api param 'usemsforovaexport' to 'forcemstodownloadvmfiles', to be generic
* Updated hardcoded UI messages with message labels
* Updated UI to support importvm api param - forcemstodownloadvmfiles
* Improved instance conversion support checks on ubuntu hosts, and for windows guest vms
* Use OVF template (VM disks and spec files) for instance conversion from VMware, instead of OVA file
- this would further increase the migration performance (as it reduces the time for OVA preparation / archiving of the VM files into a single file)
* OVF export tool parallel threads code improvements
* Updated 'convert.vmware.instance.to.kvm.timeout' config default value to 3 hrs
* Config values check & code improvements
* Updated import log, with time taken and vm details
* Support for parallel downloads of VMware VM disk files while exporting OVF from MS, and other changes below.
- Skip clone for powered off VMs
- Fixes to support standalone host (with its default datacenter)
- Some code improvements
* rebase fixes
* rebase fixes
* minor improvement
* code improvements - threads configuration, and api parameter changes to import vm files
* typo fix in error msg
* Ability to specify NFS mount options while adding a primary storage and modify it later
* Pull 8947: Rename all occurrence of nfsopt to nfsMountOpt and added nfsMountOpts to ApiConstants
* Pull 8947: Refactor code - move into separate methods
* Pull 8947: CollectionsUtils.isNotEmpty and switch statement in LibvirtStoragePoolDef.java
* Pull 8947: UI - cancel maintainenace will remount the storage pool and apply the options
* Pull 8947: UI - moved edit NFS mount options to edit Primary Storage form
* Pull 8947: UI - moved 'NFS Mount Options' to below 'Type' in dataview
* Pull 8947: Fixed message in AddPrimaryStorage.vue
* Pull 8947: Convert _nfsmountOpts to Set in libvirtStoragePoolDef
* Pull 8947: Throw exception and log error if mount fails due to incorrect mount option
* Pull 8947: Added UT and moved integration test to component/maint
* Pull 8947: Review comments
* Pull 8947: Removed password from integration test
* Pull 8947: move details allocation to inside the if loop in getStoragePoolNFSMountOpts
* Pull 8947: Fixed a bug in AddPrimaryStorage.vue
* Pull 8947: Pool should remain in maintenance mode if mount fails
* Pull 8947: Removed password from integration test
* Pull 8947: Added UT
* Pull 8875: Fixed a bug in CloudStackPrimaryDataStoreLifeCycleImplTest
* Pull 8875: Fixed a bug in LibvirtStoragePoolDefTest
* Pull 8947: minor code restructuring
* Pull 8947 : added some ut for coverage
* Fix LibvirtStorageAdapterTest UT
* Updates to change PUre and Primera to host-centric vlun assignments; various small bug fixes
* update to add timestamp when deleting pure volumes to avoid future conflicts
* update to migrate to properly check disk offering is valid for the target storage pool
* Updates to change PUre and Primera to host-centric vlun assignments; various small bug fixes
* update to add timestamp when deleting pure volumes to avoid future conflicts
* update to migrate to properly check disk offering is valid for the target storage pool
* improve error handling when copying volumes to add precision to which step failed
* rename pure volume before delete to avoid conflicts if the same name is used before its expunged on the array
* remove dead code in AdaptiveDataStoreLifeCycleImpl.java
* Fix issues found in PR checks
* fix session refresh TTL logic
* updates from PR comments
* logic to delete by path ONLY on supported OUI
* fix to StorageSystemDataMotionStrategy compile error
* change noisy debug message to trace message
* fix double callback call in handleVolumeMigrationFromNonManagedStorageToManagedStorage
* fix for flash array delete error
* fix typo in StorageSystemDataMotionStrategy
* change copyVolume to use writeback to speed up copy ops
* remove returning PrimaryStorageDownloadAnswer when connectPhysicalDisk returns false during KVMStorageProcessor template copy
* remove change to only set UUID on snapshot if it is a vmSnapshot
* reverting change to UserVmManagerImpl.configureCustomRootDiskSize
* add error checking/simplification per comments from @slavkap
* Update engine/storage/datamotion/src/main/java/org/apache/cloudstack/storage/motion/StorageSystemDataMotionStrategy.java
Co-authored-by: Suresh Kumar Anaparti <sureshkumar.anaparti@gmail.com>
* address PR comments from @sureshanaparti
---------
Co-authored-by: GLOVER RENE <rg9975@cs419-mgmtserver.rg9975nprd.app.ecp.att.com>
Co-authored-by: Suresh Kumar Anaparti <sureshkumar.anaparti@gmail.com>
* Added timeout config to copy the disks of remote KVM instance while importing the instance from an external host
* Updated copy config units to mins
* Cleanup remote converted file and local file when copy failed
* kvm: replace ISO path in vm XML configuration during vm migration
* Update 9212: address comments
* kvm: fix vm migration if there are multiple image stores
This PR addresses the issue #8789
The original issue is disconnectPhysicalDiskByPath() implementation in FibreChannelAdaptor always returns true irrespective of the success of the operation. This was already fixed in the PR #8889 .
Ideally this method has to be called after choosing the right adapter based on the storage pool type of the volume path, but currently it is just called in a loop.
05b9b6e2e7/plugins/hypervisors/kvm/src/main/java/com/cloud/hypervisor/kvm/storage/KVMStoragePoolManager.java (L200-L212)
while trying to fix the case of running into the loop of all adapters by somehow passing the storage pool type to that caller cleanup() method but this is touching all over the code (which I fear it creates other regressions), instead I feel we can keep it the current way only since Fibrechannel adapter has already fixed.
In this PR I've added the java doc explaining the method and situation.
This fixes a limitation for arm64/aarch64 KVM hosts to correctly export
the product name via sysconfig attribute. Without this `cloud-init`
doesn't function correctly on arm64 platforms.
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Add a global setting to control whether redirection is allowed while
downloading templates and volumes
Signed-off-by: Abhishek Kumar <abhishek.mrt22@gmail.com>
Add a global setting to control whether redirection is allowed while
downloading templates and volumes
core: some changes on SimpleHttpMultiFileDownloader
similar as HttpTemplateDownloader
Signed-off-by: Abhishek Kumar <abhishek.mrt22@gmail.com>
(cherry picked from commit b1642bc3bf58ccde9f56f632b5a9fe46a3eb5356)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
* Introduced a new API checkVolumeAndRepair that allows users or admins to check and repair if any leaks observed.
Currently this is supported only for KVM
* some fixes
* Added unit tests
* addressed review comments
* add repair volume while granting access
* Changed repair parameter to accept both leaks/all
* Introduced new global setting volume.check.and.repair.before.use to do volume check and repair before VM start or volume attach operations
* Added volume check and repair changes only during VM start and volume attach operations
* Refactored the names to look similar across the code
* Some code fixes
* remove unused code
* Renamed repair values
* Fixed unit tests
* changed version
* Address review comments
* Code refactored
* used volume name in logs
* Changed the API to Async and the setting scope to storage pool
* Fixed exit value handling with check volume command
* Fixed storage scope to the setting
* Fix volume format issues
* Refactored the log messages
* Fix formatting
This PR fixes bug introduced in #8502. Timeout for script execution was set to 60 ms instead of 60s which resulted in host not getting UEFI enabled. This is a blocker for 4.19 release.
We do this by introducing a new agent parameter `agent.script.timeout` (default - 60 seconds) to use as a timeout for the script checking host's UEFI status.
We also externalize the timeout for the ReadyCommand by introducing a new global setting `ready.command.wait` (default - 60 seconds).
For ModifyStoragePoolCommand, we don't externalize the timeout to avoid confusion for the user. Since, the required timeout can vary depending on the provider in use and we are only setting the wait for default host listener for now. Instead, we reuse the global `wait` setting by dividing it by `5` making the default value of 6 minutes (1800/5 = 360s) for ModifyStoragePoolCommand.
Note: the actual time, the MS waits is twice the wait set for a Command. Check reference code below.
19250403e6/engine/orchestration/src/main/java/com/cloud/agent/manager/AgentAttache.java (L406-L442)
There are a lot of test failures due to test_vm_life_cycle.py in multiple PRs due to host not available for migration of VMs.
#8438 (comment)
#8433 (comment)
#7344 (comment)
While debugging I noticed that the hosts get stuck in Connecting state because MS is waiting for a response of the ReadyCommand from the agent. Since we take a lock on connection and disconnection, restarting the agent doesn't work. To fix this, we have to restart the MS or wait for ~1 hour (default timeout).
On the agent side, it gets stuck waiting for a response from the Script execution.
To reproduce, run smoke/test_vm_life_cycle.py (TestSecuredVmMigration test class to be specific). Once the tests are complete, you will notice that some hosts are stuck in Connecting state. And restarting the agent fails due to the named lock. Locks on DB can be checked using the below query.
SELECT *
FROM performance_schema.metadata_locks
INNER JOIN performance_schema.threads ON THREAD_ID = OWNER_THREAD_ID
WHERE PROCESSLIST_ID <> CONNECTION_ID() \G;
This PR adds a wait for the ready command and a timeout to the Script execution to ensure that the thread doesn't get stuck and the named lock from database is released.
This PR fixes a regression caused by #8465 on advanced zones, import fails with:
2024-01-10 12:13:33,234 DEBUG [o.a.c.e.o.NetworkOrchestrator] (API-Job-Executor-3:ctx-991bbe9f job-128 ctx-f49517d4) (logid:d7b8e716) Allocating nic for vm 142272e8-9e2e-407b-9d7e-e9a03b81653c in network Network {"id": 204, "name": "Isolated", "uuid": "9679fac5-e3ac-4694-a57b-beb635340f39", "networkofferingid": 10} during import
2024-01-10 12:13:33,239 ERROR [o.a.c.v.UnmanagedVMsManagerImpl] (API-Job-Executor-3:ctx-991bbe9f job-128 ctx-f49517d4) (logid:d7b8e716) Failed to import NICs while importing vm: i-2-31-VM
com.cloud.exception.InsufficientVirtualNetworkCapacityException: Unable to acquire Guest IP address for network Network {"id": 204, "name": "Isolated", "uuid": "9679fac5-e3ac-4694-a57b-beb635340f39", "networkofferingid": 10}Scope=interface com.cloud.dc.DataCenter; id=1
at org.apache.cloudstack.engine.orchestration.NetworkOrchestrator.importNic(NetworkOrchestrator.java:4582)
at org.apache.cloudstack.vm.UnmanagedVMsManagerImpl.importNic(UnmanagedVMsManagerImpl.java:859)
at org.apache.cloudstack.vm.UnmanagedVMsManagerImpl.importVirtualMachineInternal(UnmanagedVMsManagerImpl.java:1198)
at org.apache.cloudstack.vm.UnmanagedVMsManagerImpl.importUnmanagedInstanceFromHypervisor(UnmanagedVMsManagerImpl.java:1511)
at org.apache.cloudstack.vm.UnmanagedVMsManagerImpl.baseImportInstance(UnmanagedVMsManagerImpl.java:1342)
at org.apache.cloudstack.vm.UnmanagedVMsManagerImpl.importUnmanagedInstance(UnmanagedVMsManagerImpl.java:1282)
at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
Also, addresses the VNC password field set instead of a fixed string
1. Problem description
In Apache CloudStack (ACS), when a VM is deployed in a host with the KVM hypervisor, an XML file is created in the assigned host, which has a property shares that defines the weight of the VM to access the host CPU. The value of this property has no unit, and it is a relative measure to calculate how much CPU a given VM will have in the host. However, this value has a limit, which depends on the version of cgroup utilized by the host's kernel. The problem lies at the range value of shares that varies between both versions: [2, 264144] for cgroups version 1; and [1, 10000] for cgroups version 2. Currently, ACS calculates the value of shares using Equation 1, presented below, where CPU is the number of cores and speed is the CPU frequency; both specified in the VM's compute offering. Therefore, if a compute offering has, for example, 6 cores at 2 GHz, the shares value will be 12000 and an exception will be thrown by libvirt if the host utilizes cgroup v2. The second version is becoming the default one in current Linux distributions; thus, it is necessary to address this limitation.
Equation 1
shares = CPU * speed
Fixes: #6744
2. Proposed changes
To address the problem described, we propose to apply a scale conversion considering the max shares of the host. Using the same formula currently utilized by ACS, it is possible to calculate the maximum shares of a VM for a given host. In other words, using the number of cores and the nominal speed of the host's CPU as the upper limit of shares allowed to a VM. Then, this value will be scaled to the allowed interval of [1, 10000] of cgroup v2 by using a linear scale conversion.
The VM shares would be calculated as Equation 2, presented below, where VM requested shares is the requested shares value calculated using Equation 1, cgroup upper limit is fixed with a value of 10000 (cgroups v2 upper limit), and host max shares is the maximum shares value of the host, calculated using Equation 1. Using Equation 2, the only case where a VM passes the cgroup v2 limit is when the user requests more resources than the host has, which is not possible with the current implementation of ACS.
Equation 2
shares = (VM requested shares * cgroup upper limit)/host max shares
To implement the proposal, the following APIs will be updated: deployVirtualMachine, migrateVirtualMachine and scaleVirtualMachine. When a VM is being deployed, a new verification will be added to find a suitable host. The max shares of each host will be calculated, and the VM calculated shares will be verified if it does not surpass the host's value. Likewise, the migration of VMs will have a similar new verification. Lastly, the scale of VMs will also have the same verification for the VM's host.
To determine the max shares of a given host, we will use the same equation currently used in ACS for calculating the shares of VMs, presented in Section 1. When Equation 1 is used to determine the maximum shares of a host, CPU is the number of cores of the host, and speed is the nominal CPU speed, i.e., considering the CPU's base frequency.
It is important to note that these changes are only for hosts with the KVM hypervisor using cgroup v2 for now.
This PR provides a new primary storage volume type called "FiberChannel" that allows access to volumes connected to hosts over fiber channel connections. It requires Multipath to provide path discovery and failover. Second, the PR adds an AdaptivePrimaryDatastoreProvider that abstracts how volumes are managed/orchestrated from the connector to communicate with the primary storage provider, using a ProviderAdapter interface, allowing the code interacting with the primary storage provider API's to be simpler and have no direct dependencies on Cloudstack code. Lastly, the PR provides an implementation of the ProviderAdapter classes for the HP Enterprise Primera line of storage solutions and the Pure Flash Array line of storage solutions.
