From current AWS documentation:
https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/volume_limits.html
T3, C5, C5d, M5, M5d, R5, R5d, and z1d instances support a maximum of
28 attachments, and every instance has at least one network interface
attachment. If you have no additional network interface attachments on
these instances, you could attach 27 EBS volumes.
Automatic merge from submit-queue (batch tested with PRs 67745, 67432, 67569, 67825, 67943). If you want to cherry-pick this change to another branch, please follow the instructions here: https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md.
Move volume dynamic provisioning scheduling to beta
**What this PR does / why we need it**:
* Combine feature gate VolumeScheduling and DynamicProvisioningScheduling into one
* Add allowedTopologies description in kubectl
**Special notes for your reviewer**:
Wait until related e2e and downside plugins are ready.
/hold
**Release note**:
```release-note
Move volume dynamic provisioning scheduling to beta (ACTION REQUIRED: The DynamicProvisioningScheduling alpha feature gate has been removed. The VolumeScheduling beta feature gate is still required for this feature)
```
Automatic merge from submit-queue (batch tested with PRs 67745, 67432, 67569, 67825, 67943). If you want to cherry-pick this change to another branch, please follow the instructions here: https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md.
Fix panic when choosing zone or zones for volume
**What this PR does / why we need it**:
Fix panic when choosing zone or zones for volume, so that zoneSlice won't divide by zero now.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Fixes #
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
cc @ddebroy @andyzhangx
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Attacher/Detacher refactor for local storage
Proposal link: https://github.com/kubernetes/community/pull/2438
**What this PR does / why we need it**:
Attacher/Detacher refactor for the plugins which just need to mount device, but do not need to attach, such as local storage plugin.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Fixes #
**Special notes for your reviewer**:
```release-note
Attacher/Detacher refactor for local storage
```
/sig storage
/kind feature
Automatic merge from submit-queue (batch tested with PRs 67396, 67097, 67395, 67365, 67099). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Ignore EIO error in unmount path
**What this PR does / why we need it**:
This PR ignores EIO in unmount path. XFS shuts down filesystem when the target is down and it returns EIO for the stat calls used in unmount path.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Fixes#66868
**Special notes for your reviewer**:
We already handle ESTALE & ENOTCONN errors in isCorruptedMnt Call. Adding EIO to that list covers the XFS shutdown case.
Also Flexvolume doesn't check for these errors in its current form. Updated Flexvolume code to handle it.
```release-note
NONE
```
There are two motivations for this change:
(1) CSI plugins are soon going to support volume expansion. For such
plugins, admission controller doesn't know whether the plugins are
capabale of supporting volume expansion or not.
(2) Currently, admission controller rejects PVC updates for in-tree plugins
that don't support volume expansion (e.g., NFS, iSCSI). This change allows
external controllers to expand volumes similar to how external provisioners
operate.
Automatic merge from submit-queue (batch tested with PRs 64844, 63176). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Fix discovery/deletion of iscsi block devices
This PR modifies the iSCSI attach/detatch codepaths in the following
ways:
1) After unmounting a filesystem on an iSCSI block device, always
flush the multipath device mapper entry (if it exists) and delete
all block devices so the kernel forgets about them.
2) When attaching an iSCSI block device, instead of blindly
attempting to scan for the new LUN, first determine if the target
is already logged into, and if not, do the login first. Once every
portal is logged into, the scan is done.
3) Scans are now done for specific devices, instead of the whole
bus. This avoids discovering LUNs that kubelet has no interest in.
4) Additions to the underlying utility interfaces, with new tests
for the new functionality.
5) Some existing code was shifted up or down, to make the new logic
work.
6) A typo in an existing exec call on the attach path was fixed.
Fixes#59946
```release-note
When attaching iSCSI volumes, kubelet now scans only the specific
LUNs being attached, and also deletes them after detaching. This avoids
dangling references to LUNs that no longer exist, which used to be the
cause of random I/O errors/timeouts in kernel logs, slowdowns during
block-device related operations, and very rare cases of data corruption.
```
This change ensures that iSCSI block devices are deleted after
unmounting, and implements scanning of individual LUNs rather
than scanning the whole iSCSI bus.
In cases where an iSCSI bus is in use by more than one attachment,
detaching used to leave behind phantom block devices, which could
cause I/O errors, long timeouts, or even corruption in the case
when the underlying LUN number was recycled. This change makes
sure to flush references to the block devices after unmounting.
The original iSCSI code scanned the whole target every time a LUN
was attached. On storage controllers that export multiple LUNs on
the same target IQN, this led to a situation where nodes would
see SCSI disks that they weren't supposed to -- possibly dozens or
hundreds of extra SCSI disks. This caused 3 significant problems:
1) The large number of disks wasted resources on the node and
caused a minor drag on performance.
2) The scanning of all the devices caused a huge number of uevents
from the kernel, causing udev to bog down for multiple minutes in
some cases, triggering timeouts and other transient failures.
3) Because Kubernetes was not tracking all the "extra" LUNs that
got discovered, they would not get cleaned up until the last LUN
on a particular target was detached, causing a logout. This led
to significant complications:
In the time window between when a LUN was unintentially scanned,
and when it was removed due to a logout, if it was deleted on the
backend, a phantom reference remained on the node. In the best
case, the phantom LUN would cause I/O errors and timeouts in the
udev system. In the worst case, the backend could reuse the LUN
number for a new volume, and if that new volume were to be
scheduled to a pod with a phantom reference to the old LUN by the
same number, the initiator could get confused and possibly corrupt
data on that volume.
To avoid these problems, the new implementation only scans for
the specific LUN number it expects to see. It's worth noting that
the default behavior of iscsiadm is to automatically scan the
whole bus on login. That behavior can be disabled by setting
node.session.scan = manual
in iscsid.conf, and for the reasons mentioned above, it is
strongly recommended to set that option. This change still works
regardless of the setting in iscsid.conf, and while automatic
scanning will cause some problems, this change doesn't make the
problems any worse, and can make things better in some cases.
Automatic merge from submit-queue (batch tested with PRs 65492, 65516, 65447). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Resolve potential devicePath symlink when MapVolume in containerized kubelet
**What this PR does / why we need it**: Ensures local block volumes will work in case kubelet is running in a container
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Fixes#65445
**Special notes for your reviewer**: Code is mostly plumbing. If there is a better way to do it, let me know :)
I assume there will be e2e tests for the non-containerized case. I will need to test the containerized case myself, which may take a while.
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 64276, 64094, 64719, 64766, 64750). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Delegate map operation to BlockVolumeMapper plugin
**What this PR does / why we need it**:
This PR refactors the volume controller's operation generator, for block mapping, to delegate core block mounting sequence to the `volume.BlockVolumeMapper` plugin instead of living in the operation generator. This is to ensure better customization of block volume logic for existing internal volume plugins.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Fixes#64093
```release-note
NONE
```
/sig storage
Automatic merge from submit-queue (batch tested with PRs 62266, 64351, 64366, 64235, 64560). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Add log and fs stats for Windows containers
**What this PR does / why we need it**:
Add log and fs stats for Windows containers.
Without this, kubelet will report errors continuously:
```
Unable to fetch container log stats for path \var\log\pods\2a70ed65-37ae-11e8-8730-000d3a14b1a0\echo: Du not supported for this build.
```
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Fixes#60180#62047
**Special notes for your reviewer**:
**Release note**:
```release-note
Add log and fs stats for Windows containers
```
Automatic merge from submit-queue (batch tested with PRs 63348, 63839, 63143, 64447, 64567). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Add block volume support to internal provisioners.
**What this PR does / why we need it**:
Internal provisioners now create filesystem PVs when block PVs are requested. This leads to unbindable PVCs.
In this PR, volume plugins that support block volumes provision block PVs when block is requested. All the other provisioners return clear error in `kubectl describe pvc`:
```
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Warning ProvisioningFailed 7s (x2 over 18s) persistentvolume-controller Failed to provision volume with StorageClass "standard": kubernetes.io/cinder does not support block volume provisioning
```
```release-note
NONE
```
cc @kubernetes/vmware for vsphere changes
cc @andyzhangx for Azure changes
/assign @copejon @mtanino
Automatic merge from submit-queue (batch tested with PRs 59034, 63565, 63533). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
[RBD block device ]ignore the loopbackdevice error, or the rbd volume will not get detached
The rbd volume loop device maybe not found and then the volume will not get detached.
so need to ignore the error.
@mtanino @sbezverk
Can you take a look?
Release note:
```
ignore the loopback device not found error when `GetLoopDevice`
```
Automatic merge from submit-queue (batch tested with PRs 62650, 62303, 62545, 62375). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Fix user visible files creation for windows
**What this PR does / why we need it**:
Fix user visible files creation for windows. Without this, [createUserVisibleFiles](https://github.com/kubernetes/kubernetes/blob/master/pkg/volume/util/atomic_writer.go#L415:24) will get linkname with subpath included, and then symlink will fail. This is because "/" is used in pod spec (e.g. `"new/path/data-1"`) while "\" is used on Windows to get linkname.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Fixes#62338
**Special notes for your reviewer**:
Should also be cherry-picked to old releases.
**Release note**:
```release-note
Fix user visible files creation for windows
```
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Fix volume node affinity to OR node selector terms
**What this PR does / why we need it**:
Fixes node selector terms to be ORed, to be consistent with documentation and Pod.NodeAffinity. Also handles the "node selector term nil or empty matches nothing" behavior.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Fixes#62551
**Special notes for your reviewer**:
**Release note**:
```release-note
Fixes issue where PersistentVolume.NodeAffinity.NodeSelectorTerms were ANDed instead of ORed.
```
In WatchPod(), if one of the two channels being watched (pod updates and
events) is closed, the for/select loop turns into a tight infinite loop because
the select immediately falls through due to the channel being closed. Watch
them independently instead.
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Add volume spec to mountedPod in actual state of world
Add volume spec into mountedPod data struct in the actual state of the
world.
Fixes issue #61248
use MakeAbsolutePath to convert path in Windows
fix test error: allow relative path for local volume
fix comments
fix comments and add windows unit tests
Automatic merge from submit-queue (batch tested with PRs 61894, 61369). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Use range in loops; misc fixes
**What this PR does / why we need it**:
It is cleaner to use `range` in for loops to iterate over channel until it is closed.
**Release note**:
```release-note
NONE
```
/kind cleanup
David Zhu