rbd mirroring CLI calls are async and it doesn't wait
for the operation to be completed. ex:- `rbd mirror image enable`
it will enable the mirroring on the image but it doesn't
ensure that the image is mirroring enabled and healthy
primary. The same goes for the promote volume also.
This commits adds a check-in PromoteVolume to make sure
the image in a healthy state i.e `up+stopped`.
note:- not considering any intermediate states to make
sure the image is completely healthy before responding
success to the RPC call.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Journal-based RADOS block device mirroring ensures point-in-time
consistent replicas of all changes to an image, including reads and
writes, block device resizing, snapshots, clones, and flattening.
Journaling-based mirroring records all modifications to an image in the
order in which they occur. This ensures that a crash-consistent mirror
of an image is available.
Mirroring when configured in journal mode, mirroring will
utilize the RBD journaling image feature to replicate the image
contents. If the RBD journaling image feature is not yet enabled on the
image, it will be automatically enabled.
Fixes: #2018
Co-authored-by: Madhu Rajanna <madhupr007@gmail.com>
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
Depending on the way Ceph-CSI is deployed, the capabilities will be
configured for the GetCapabilities procedure. The other procedures are
more straight-forward.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
After adding the new CSI-Addons Server, golang-ci complains that
driver.Run() is too complex. By moving the profiling checks and starting
of the go-routines in their own function, golang-ci is happy again.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
currently we are fist operating on the dummy
image to refresh the pool and then we are adding
the scheduling. we think the scheduling should
be added first and than we should refresh the
pool. If we do this all the existing schedules
will be considered from the scheduler.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
with shallow copy of rbdVol to dummyVol
the image name update of the dummyVol is getting
reflected on the rbdVol which we dont want.
do deep copy to avoid this problem.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Uses the below schema to supply mounter specific map/unmapOptions to the
nodeplugin based on the discussion we all had at
https://github.com/ceph/ceph-csi/pull/2636
This should specifically be really helpful with the `tryOthermonters`
set to true, i.e with fallback mechanism settings turned ON.
mapOption: "kbrd:v1,v2,v3;nbd:v1,v2,v3"
- By omitting `krbd:` or `nbd:`, the option(s) apply to
rbdDefaultMounter which is krbd.
- A user can _override_ the options for a mounter by specifying `krbd:`
or `nbd:`.
mapOption: "v1,v2,v3;nbd:v1,v2,v3"
is effectively the same as the 1st example.
- Sections are split by `;`.
- If users want to specify common options for both `krbd` and `nbd`,
they should mention them twice.
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
The dummy mirror image needs to be disabled and then
reenabled for mirroring, to ensure a newly promoted
primary is now starting to schedule snapshots.
Signed-off-by: Shyamsundar Ranganathan <srangana@redhat.com>
currently we have a bug in rbd mirror scheduling module.
After doing failover and failback the scheduling is not
getting updated and the mirroring snapshots are not
getting created periodically as per the scheduling
interval. This PR workarounds this one by doing below
operations
* Create a dummy (unique) image per cluster and this image
should be easily identified.
* During Promote operation on any image enable the
mirroring on the dummy image. when we enable the mirroring
on the dummy image the pool will get updated and the
scheduling will be reconfigured.
* During Demote operation on any image disable the mirroring
on the dummy image. the disable need to be done to enable
the mirroring again when we get the promote request to make
the image as primary
* When the DR is no more needed, this image need to be
manually cleanup as for now as we dont want to add a check
in the existing DeleteVolume code path for delete dummy image
as it impact the performance of the DeleteVolume workflow.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Moved to add scheduling to the promote
operation as scheduling need to be added
when the image is promoted and this is
the correct method of adding the scheduling
to make the scheduling take place.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
instead of logging the volumeID and the pool
name. log the poolname and image name for better
debugging.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
On line 341 a `transaction` is created. This is passed to the deferred
`undoStagingTransaction()` function when an error in the
`NodeStageVolume` procedure is detected. So far, so good.
However, on line 356 a new `transaction` is returned. This new
`transaction` is not used for the defer call.
By removing the empty `transaction` that is used in the defer call, and
calling `undoStagingTransaction()` on an error of `stageTransaction()`,
the code is a little simpler, and the cleanup of the transaction should
be done correctly now.
Updates: #2610
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Problem:
On remap/attach of device (i.e. nodeplugin restart), there is no way
for rbd-nbd to defend if the backend storage is matching with the initial
backend storage.
Say, if an initial map request for backend "pool1/image1" got mapped to
/dev/nbd0 and the userspace process is terminated (on nodeplugin restart).
A next remap/attach (nodeplugin start) request within reattach-timeout is
allowed to use /dev/nbd0 for a different backend "pool1/image2"
For example, an operation like below could be dangerous:
$ sudo rbd-nbd map --try-netlink rbd-pool/ext4-image
/dev/nbd0
$ sudo blkid /dev/nbd0
/dev/nbd0: UUID="bfc444b4-64b1-418f-8b36-6e0d170cfc04" TYPE="ext4"
$ sudo pkill -15 rbd-nbd <-- nodeplugin terminate
$ sudo rbd-nbd attach --try-netlink --device /dev/nbd0 rbd-pool/xfs-image
/dev/nbd0
$ sudo blkid /dev/nbd0
/dev/nbd0: UUID="d29bf343-6570-4069-a9ea-2fa156ced908" TYPE="xfs"
Solution:
rbd-nbd/kernel now provides a way to keep some metadata in sysfs to identify
between the device and the backend, so that when a remap/attach request is
made, rbd-nbd can compare and avoid such dangerous operations.
With the provided solution, as part of the initial map request, backend
cookie (ceph-csi VOLID) can be stored in the sysfs per device config, so
that on a remap/attach request rbd-nbd will check and validate if the
backend per device cookie matches with the initial map backend with the help
of cookie.
At Ceph-csi we use VOLID as device cookie, which will be unique, we pass
the VOLID as cookie at map and use the same at the time of attach, that
way rbd-nbd can identify backends and their matching devices.
Requires:
https://github.com/ceph/ceph/pull/41323https://lkml.org/lkml/2021/4/29/274
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
This change allows the user to choose not to fallback to NBD mounter
when some ImageFeatures are absent with krbd driver, rather just fail
the NodeStage call.
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
Currently, we recognize and warn for the provided image features based on
our prior intelligence at ceph-csi (i.e based on supportedFeatures map
and validateImageFeatures) at image/PV creation time. It might be very
much possible that the cluster is heterogeneous i.e. the PV creation and
application container might both be on different nodes with different
kernel versions (krbd driver versions).
This PR adds a mechanism to check for the supported krbd features during
mount time, if the krbd driver doesn't have the specified image feature
then it will fall back to rbd-nbd mounter.
Fixes: #478
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
this commit make use of the migration request secret parsing and set
the required fields for further nodestage operations
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
parseAndDeleteMigratedVolume() prviously clubbed the logic of
parsing of migration volume handle and then continued with the
deletion of the volume. however this commit split this
logic into two, ie parsing has been done in parseMigrationVolID()
and DeleteMigratedVolume() deletes the backend volume.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
This commit adds a couple of helper functions to parse the migration
request secret and set it for further csi driver operations.
More details:
The intree secret has a data field called "key" which is the base64
admin secret key. The ceph CSI driver currently expect the secret to
contain data field "UserKey" for the equivalant. The CSI driver also
expect the "UserID" field which is not available in the in-tree secret
by deafult. This missing userID will be filled (if the username differ
than 'admin') in the migration secret as 'adminId' field in the
migration request, this commit adds the logic to parse this migration
secret as below:
"key" field value will be picked up from the migraion secret to "UserKey"
field.
"adminId" field value will be picked up from the migration secret to "UserID"
field
if `adminId` field is nil or not set, `UserID` field will be filled with
default value ie `admin`.The above logic get activated only when the secret
is a migration secret, otherwise skipped to the normal workflow as we have
today.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
Thick-provisioning was introduced to make accounting of assigned space
for volumes easier. When thick-provisioned volumes are the only consumer
of the Ceph cluster, this works fine. However, it is unlikely that this
is the case. Instead, accounting of the requested (thin-provisioned)
size of volumes is much more practical as different types of volumes can
be tracked.
OpenShift already provides cluster-wide quotas, which can combine
accounting of requested volumes by grouping different StorageClasses.
In addition to the difficult practise of allowing only thick-provisioned
RBD backed volumes, the performance makes thick-provisioning
troublesome. As volumes need to be completely allocated, data needs to
be written to the volume. This can take a long time, depending on the
size of the volume. Provisioning, cloning and snapshotting becomes very
much noticeable, and because of the additional time consumption, more
prone to failures.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
for comparing the image states use the states
defined in the go-ceph avoid creating of the
deplicate const in cephcsi.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
below are the local states of the mirrored image
"unknown" -> If the image is in an error state
means data is completely synced
"error" -> If the image is in an error state
means it needs resync
"syncing"
"starting_replay"
"replaying"
"stopping_replay"
"stopped"
If the resync is successfully started which
means the image will be in "replaying" state.
we can consider "replaying" state to report
resync succesfully going on state.
we are discarding the intermediate states like
"syncing", "starting_replay" and "stopping_replay".
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
After moving moving image to trash, if `trash remove` step fails,
then external-provisioner will issue subsequent requests, in which
image will be absent in pool( will be in trash) and omap cleanup will
be done with stale image left in trash with no `trash remove` step on it.
To avoid this scenario list trash images and find corresponding id for given
image name and add a task to flatten when we encounter a ErrImageNotFound.
Fixes: #1728
Signed-off-by: Rakshith R <rar@redhat.com>
In some corner case like `re-player shutdown` the
local image will not be in error state. It would
be also worth considering `description` field to
make sure about split-brain.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
previously we were retriving clusterID using the monitors field
in the volume context at node stage code path. however it is possible to
retrieve or use clusterID directly from the volume context. This
commit also remove the getClusterIDFromMigrationVolume() function
which was used previously and its tests
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
we reuse or overload the variable name in the test execution at present.
This commit use a different variable name as initialized in each run
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
For static volume, the user will manually mounts
already existing image as a volume to the application
pods. As its a rbd Image, if the PVC is of type
fileSystem the image will be mapped, formatted
and mounted on the node,
If the user resizes the image on the ceph cluster.
User cannot not automatically resize the filesystem
created on the rbd image. Even if deletes and
recreates the kubernetes objects, the new size
will not be visible on the node.
With this changes During the NodeStageVolumeRequest
the nodeplugin will check the size of the mapped rbd
image on the node using the devicePath. and also
the rbd image size on the ceph cluster.
If the size is not matching it will do the file
system resize on the node as part of the
NodeStageVolumeRequest RPC call.
The user need to do below operation to see new size
* Resize the rbd image in ceph cluster
* Scale down all the application pods using the static
PVC.
* Make sure no application pods which are using the
static PVC is running on a node.
* Scale up all the application pods.
Validate the new size in application pod mounted
volume.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
in NodeStage operation we are flattening
the image to support mounting on the older
clients. this commits moves it to a helper
function to reduce code complexity.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
During PVC snapshot/clone both kms config and passphrase needs to copied,
while for PVC restore only passphrase needs to be copied to dest rbdvol
since destination storageclass may have another kms config.
Signed-off-by: Rakshith R <rar@redhat.com>
This commit adds the logic to detect a passed in volumeID
is a migrated volume ID and if yes, the driver connect to the
backend cluster and clean/delete the image. The logic
only applied if its a migration volume ID. The migration volume ID
carry the information like mons, pool and image name which is
good enough for the driver to identify and connect to the backend
cluster for its operations.
migration volID format:
<mig>_mons-<monsHash>_image-<imageUID>_<poolHash>
Details on the hash values:
* MonsHash: this carry a hash value (md5sum) which will be acted as the
`clusterID` for the operations in this context.
* ImageUID: this is the unique UUID generated by kubernetes for the created
volume.
* PoolHash: this is an encoded string of pool name.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
as we are refractoring the cephfs code,
Moving all the core functions to a new folder
/pkg called core. This will make things easier
to implement. For now onwards all the core
functionalities will be added to the core
package.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
the migration nodestage request does not carry the 'clusterID' in it
and only monitors are available with the volumeContext. The volume
context flag 'migration=true' and 'static=true' flags allow us to
fill 'clusterID' from the passed in monitors to the volume Context,so
that rest of the static operations on nodestage can be proceeded as we
do treat static volumes today.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
Currently, we delete the ceph client log file on unmap/detach.
This patch provides additional alternatives for users who would like to
persist the log files.
Strategies:
-----------
`remove`: delete log file on unmap/detach
`compress`: compress the log file to gzip on unmap/detach
`preserve`: preserve the log file in text format
Note that the default strategy will be remove on unmap, and these options
can be tweaked from the storage class
Compression size details example:
On Map: (with debug-rbd=20)
---------
$ ls -lh
-rw-r--r-- 1 root root 526K Sep 1 18:15
rbd-nbd-0001-0024-fed5480a-f00f-417a-a51d-31d8a8144c03-0000000000000003-d2e89c87-0b4d-11ec-8ea6-160f128e682d.log
On unmap:
---------
$ ls -lh
-rw-r--r-- 1 root root 33K Sep 1 18:15
rbd-nbd-0001-0024-fed5480a-f00f-417a-a51d-31d8a8144c03-0000000000000003-d2e89c87-0b4d-11ec-8ea6-160f128e682d.gz
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
Log:
internal/rbd/rbd_attach.go:424:2: hugeParam: dArgs is heavy (88 bytes);
consider passing it by pointer (gocritic)
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
Currently we return a !ready status if an image
is not found when a replication resync is issued.
We also return a !ready just post issuing a resync.
The change is to ensure we return errors in these
cases for the caller to retry the operation till
we can determine we are actually resyncing, and then
return !ready with nil errors.
Part of addressing:
https://github.com/csi-addons/volume-replication-operator/issues/101
Signed-off-by: Shyamsundar Ranganathan <srangana@redhat.com>
This commit:
- modifies GetMonsAndClusterID() to take clusterID instead of options.
- moves out validation of clusterID is set or not out of GetMonsAndClusterID().
- defines ErrClusterIDNotSet new error for reusability.
- add GetClusterID() to obtain clusterID from options.
Signed-off-by: Rakshith R <rar@redhat.com>
This commit adds capability to genVolFromVolumeOptions() to fetch
mapped clusted-id & mon ips for mirrored PVC on secondary cluster
which may have different cluster-id.
This is required for NodeStageVolume().
We also don't need to check for mapping during volume create requests,
so it can be disabled by passing a bool checkClusterIDMapping as false.
GetMonsAndClusterID() is modified to accept bool checkClusterIDMapping
based on which clustermapping is checked to fetch mapped cluster-id and
mon-ips.
Signed-off-by: Rakshith R <rar@redhat.com>
This commit calls WriteCephConfig() in cephcsi.go to
create ceph.conf and keyring if it is not mounted to
be used by all cli calls and conn cmds.
Before this change, rbd-controller/omap-generator did not create
ceph.conf on startup.
Signed-off-by: Rakshith R <rar@redhat.com>
Thanks to the random unmap failure on my local machine:
I0901 17:08:37.841890 2617035 cephcmds.go:55] ID: 11 Req-ID:
0001-0024-fed5480a-f00f-417a-a51d-31d8a8144c03-0000000000000003-024983f3-0b47-11ec-8fcb-e671f0b9f58e
an error (exit status 22) occurred while running rbd args: [unmap
rbd-pool/csi-vol-024983f3-0b47-11ec-8fcb-e671f0b9f58e --device-type nbd
--options try-netlink --options reattach-timeout=300 --options
io-timeout=0]
Noticed the map args are also getting passed to/as unmap args, which is not
correct. We have separate things for mapOptions and unmapOptions. This PR
makes sure that the map args are not passed at the time of unmap.
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
At present the nodeStageVolume() handle many logic of filling rbdvol
struct based on the request received and this method is complex to
follow. with this patch, filling or populating volOptions has been
segregrated and handled hence make the stage functions' job easy.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
MirrorImageState (type C.rbd_mirror_image_state_t) has a string
method which can be used while returning error in the replication
controller. Previously, we were using int return in the error which
is not the proper usage.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
all the error check scenarios of genVolFromVolID() and unreserving
omap entries based on the error made deleteVolume method complex,
this patch create a new function which handle the error check and
unrerving omap entries accordingly and finally return the response
to deletevolume/caller.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
When NewK8sClient() detects and error, it used to call FatalLogMsg()
which causes a panic. There are additional features that can be used on
Kubernetes clusters, but these are not a requirement for most
functionalities of the driver.
Instead of causing a panic, returning an error should suffice. This
allows using the driver on non-Kubernetes clusters again.
Fixes: #2452
Signed-off-by: Niels de Vos <ndevos@redhat.com>
This commit adds fetchMappedClusterIDAndMons() which returns
monitors and clusterID info after checking cluster mapping info.
This is required for regenerating omap entries in mirrored cluster
with different clusterID.
Signed-off-by: Rakshith R <rar@redhat.com>
This commit moves getMappedID() from rbd to util
package since it is not rbd specific and exports
it from there.
Signed-off-by: Rakshith R <rar@redhat.com>
A new "internal/kms" package is introduced, it holds the API that can be
consumed by the RBD components.
The KMS providers are currently in the same package as the API. With
later follow-up changes the providers will be placed in their own
sub-package.
Because of the name of the package "kms", the types, functions and
structs inside the package should not be prefixed with KMS anymore:
internal/kms/kms.go:213:6: type name will be used as kms.KMSInitializerArgs by other packages, and that stutters; consider calling this InitializerArgs (golint)
Updates: #852
Signed-off-by: Niels de Vos <ndevos@redhat.com>
By placing the NewK8sClient() function in its own package, the KMS API
can be split from the "internal/util" package. Some of the KMS providers
use the NewK8sClient() function, and this causes circular dependencies
between "internal/utils" -> "internal/kms" -> "internal/utils", which
are not alowed in Go.
Updates: #852
Signed-off-by: Niels de Vos <ndevos@redhat.com>
if the volumeattachment has been fetched but marked for deletion
the nbd healer dont want to process further on this pv. This patch
adds a check for pv is marked for deletion and if so, make the
healer skip processing the same
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
Moving the log functions into its own internal/util/log package makes it
possible to split out the humongous internal/util packages in further
smaller pieces. This reduces the inter-dependencies between utility
functions and components, preventing circular dependencies which are not
allowed in Go.
Updates: #852
Signed-off-by: Niels de Vos <ndevos@redhat.com>
With the tests at CI, it kind of looks like that the IO is timing out after
30 seconds (default with rbd-nbd). Since we have tweaked reattach-timeout
to 300 seconds at ceph-csi, we need to explicitly set io-timeout on the
device too, as it doesn't make any sense to keep
io-timeout < reattach-timeout
Hence we set io-timeout for rbd nbd to 0. Specifying io-timeout 0 tells
the nbd driver to not abort the request and instead see if it can be
restarted on another socket.
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
Suggested-by: Ilya Dryomov <idryomov@redhat.com>
- Update the meta stash with logDir details
- Use the same to remove logfile on unstage/unmap to be space efficient
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
- One logfile per device/volume
- Add ability to customize the logdir, default: /var/log/ceph
Note: if user customizes the hostpath to something else other than default
/var/log/ceph, then it is his responsibility to update the `cephLogDir`
in storageclass to reflect the same with daemon:
```
cephLogDir: "/var/log/mynewpath"
```
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
If the image is in a secondary state and its
up+replaying means its an healthy secondary
and the image is primary somewhere in the remote cluster
and the local image is getting replayed. Delete the
OMAP data generated as we cannot delete the
secondary image. When the image on the primary
cluster gets deleted/mirroring disabled, the image on
all the remote (secondary) clusters will get
auto-deleted. This helps in garbage collecting
the OMAP, PVC and PV objects after failback operation.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
If the image is in secondary state and its
up+replaying means its an healthy secondary
and the image is primary somewhere in the remote
cluster and the local image is getting replayed.
Return success for the Disabling mirroring as
we cannot disable the mirroring on the secondary
state, when the image on the remote site gets
disabled the image on all the remote (secondary)
will get auto deleted. This helps in garbage
collecting the volume replication kuberentes
artifacts
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
This commit adds functionality of extracting encryption kmsID,
owner from volumeAttributes in RegenerateJournal() and adds utility
functions ParseEncryptionOpts and FetchEncryptionKMSID.
Signed-off-by: Rakshith R <rar@redhat.com>
This commit refractors RegenerateJournal() to take in
volumeAttributes map[string]string as argument so it
can extract required attributes internally.
Signed-off-by: Rakshith R <rar@redhat.com>
consider the empty mirroring mode when
validating the snapshot interval and
the scheduling time.
Even if the mirroring Mode is not set
validate the snapshot scheduling details
as cephcsi sets the mirroring mode to default
snapshot.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
This commit fixes snapshot id idempotency issue by
always returning an error when flattening is in progress
and not using `readyToUse:false` response.
Signed-off-by: Rakshith R <rar@redhat.com>
This commit fixes a bug in checkCloneImage() which was caused
by checking cloned image before checking on temp-clone image snap
in a subsequent request which lead to stale images. This was solved
by checking temp-clone image snap and flattening temp-clone if
needed.
This commit also fixes comparison bug in flattenCloneImage().
Signed-off-by: Rakshith R <rar@redhat.com>
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Volume generated from snap using genrateVolFromSnap
already copies volume ID correctly, therefore removing
`vol.VolID = rbdVol.VolID` which wrongly copies parent
Volume ID instead leading to error from copyEncryption()
on parent and clone volume ID being equal.
Signed-off-by: Rakshith R <rar@redhat.com>
Whenever Ceph-CSI receives a CSI/Replication
request it will first decode the
volumeHandle and try to get the required
OMAP details if it is not able to
retrieve, receives a `Not Found` error
message and Ceph-CSI will check for the
clusterID mapping. If the old volumeID
`0001-00013-site1-storage-0000000000000001
-b0285c97-a0ce-11eb-8c66-0242ac110002`
contains the `site1-storage` as the clusterID,
now Ceph-CSI will look for the corresponding
clusterID `site2-storage` from the above configmap.
If the clusterID mapping is found now Ceph-CSI
will look for the poolID mapping ie mapping between
`1` and `2`. Example:- pool with name exists on
both the clusters with different ID's Replicapool
with ID `1` on site1 and Replicapool with ID `2`
on site2. After getting the required mapping Ceph-CSI
has the required information to get more details
from the rados OMAP. If we have multiple clusterID mapping
it will loop through all the mapping and checks the
corresponding pool to get the OMAP data. If the clusterID
mapping does not exist Ceph-CSI will return an `Not Found`
error message to the caller.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Currently we have a bug that we are not using rados
namespace when adding ceph manager command to
remove the image from the trash. This commit
adds the missing rados namespace when adding
ceph manager task.
without fix the image will be moved to trash
and no task will be added to remove from the
trash. it will become ceph responsibility to
remove the image from trash when it will cleanup
the trash.
workaroud: manually purge the trash
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
RBD image metadata keys that start with '.rbd' are expected to be
internal to RBD itself and are not mirrored to remote sites. Renaming
the keys (dropping the '.' prefix) and using the new MigrateMetadata()
function now makes the keys available on remote sites too.
Closes: #2219
Signed-off-by: Niels de Vos <ndevos@redhat.com>
The new MigrateMetadata() function can be used to get the metadata of an
image with a deprecated and new key. Renaming metadata keys can be done
easily this way.
A default value will be set in the image metadata when it is missing
completely. But if the deprecated key was set, the data is stored under
the new key and the deprecated key is removed.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Currently, getImageMirroringStatus() is using RBD CLI.
This commit converts RBD CLI to go-ceph API.
Fixes: #2120
Signed-off-by: Yati Padia <ypadia@redhat.com>
Previously in ControllerExpandVolume() we had a check for encrypted
volumes and we use to fail for all expand requests on an encrypted
volume. Also for Block VolumeMode PVCs NodeExpandVolume used to be
ignored/skipped.
With these changes, we add support for the expansion of encrypted volumes.
Also for raw Block VolumeMode PVCs with Encryption we call NodeExpandVolume.
That said,
With LUKS1, cryptsetup utility doesn't prompt for a passphrase on resizing
the crypto mapper device. This is because LUKS1 devices don't use kernel
keyring for volume keys.
Whereas, LUKS2 devices use kernel keyring for volume key by default, i.e.
cryptsetup utility asks for a passphrase if it detects volume key was
previously passed to dm-crypt via kernel keyring service, we are overriding
the default by --disable-keyring option during cryptsetup open command.
So that at the time of crypto mapper device resize we will not be
prompted for any passphrase.
Fixes: #1469
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
This commit modifies the error of godot, cyclop,
paralleltest linter caused due to merged PRs.
Updates: #1586
Signed-off-by: Yati Padia <ypadia@redhat.com>
nlreturn linter requires a new line before return
and branch statements except when the return is alone
inside a statement group (such as an if statement) to
increase code clarity. This commit addresses such issues.
Updates: #1586
Signed-off-by: Rakshith R <rar@redhat.com>
This commit resolves errorlint issues
which checks for the code that will cause
problems with the error wrapping scheme.
Updates: #1586
Signed-off-by: Yati Padia <ypadia@redhat.com>
Problem:
-------
For rbd nbd userspace mounter backends, after a restart of the nodeplugin
all the mounts will start seeing IO errors. This is because, for rbd-nbd
backends there will be a userspace mount daemon running per volume, post
restart of the nodeplugin pod, there is no way to restore the daemons
back to life.
Solution:
--------
The volume healer is a one-time activity that is triggered at the startup
time of the rbd nodeplugin. It navigates through the list of volume
attachments on the node and acts accordingly.
For now, it is limited to nbd type storage only, but it is flexible and
can be extended in the future for other backend types as needed.
From a few feets above:
This solves a severe problem for nbd backed csi volumes. The healer while
going through the list of volume attachments on the node, if finds the
volume is in attached state and is of type nbd, then it will attempt to
fix the rbd-nbd volumes by sending a NodeStageVolume request with the
required volume attributes like secrets, device name, image attributes,
and etc.. which will finally help start the required rbd-nbd daemons in
the nodeplugin csi-rbdplugin container. This will allow reattaching the
backend images with the right nbd device, thus allowing the applications
to perform IO without any interruptions even after a nodeplugin restart.
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
As part of stage transaction if the mounter is of type nbd, then capture
device path after a successful rbd-nbd map.
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
Considering kubelet make sure the stage and publish operations
are serialized, we dont need any extra locking in nodePublish
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
Considering kubelet make sure the unstage and unpublish operations
are serialized, we dont need any extra locking in nodeUnpublish
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
This commit resolves parallel test issues
and also excludes internal/util/conn_pool_test.go
as those test can't run in parallel.
Updates: #1586
Signed-off-by: Yati Padia <ypadia@redhat.com>
This commit resolves godot linter issue
which says "Comment should end in a period (godot)".
Updates: #1586
Signed-off-by: Yati Padia <ypadia@redhat.com>
setting metadata in isThickProvisioned method
helps us to avoid checking thick metakey and
deprecated metakey for both thick and thin
provisioned images and also this will easily
help us to migrated the deprecated key to new key.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
isThickProvisioned is already method of the rbdImage
to keep similar thick provisioner related functions
common making isThickProvisioned as method of rbdImage.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
instead of checking the parent is thick provisioned
or not we can decide based on the rbdVol generated
from the request. If the request is to create a Thick
Image. set metadata without checking the parent.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
As image metadata key starting with '.rbd' will not
be copied when we do clone or mirroring, deprecating
the old key for the same reason use
'csi.ceph.com/thick-provisioned' to set image metadata.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Mirror-snapshots can also be automatically created on a
periodic basis if mirror-snapshot schedules are defined.
The mirror-snapshot can be scheduled globally, per-pool,
or per-image levels. Multiple mirror-snapshot schedules
can be defined at any level.
To create a mirror-snapshot schedule with rbd, specify
the mirror snapshot schedule add command along with an
optional pool or image name; interval; and optional start time:
The interval can be specified in days, hours, or minutes
using d, h, m suffix respectively. The optional start-time
can be specified using the ISO 8601 time format. For example:
```
$ rbd --cluster site-a mirror snapshot schedule
add --pool image-pool --image image1 24h 14:00:00-05:00
```
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Updated kubernetes packages to latest release.
resizefs package has been included into k8s.io/mount-utils
package. updated code to use the same.
Updates: #1968
Signed-off-by: Rakshith R <rar@redhat.com>
We have many declarations and invocations..etc with long lines which are
very difficult to follow while doing code reading. This address the issues
in below files, and restrict the line length to 120 chars.
-internal/rbd/rbd_attach.go
-internal/rbd/rbd_journal.go
-internal/rbd/rbd_util.go
-internal/rbd/replicationcontrollerserver.go
-internal/rbd/snapshot.go
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
We have many declarations and invocations..etc with long lines which are
very difficult to follow while doing code reading. This address the issues
in 'internal/rbd/*server.go' and 'internal/rbd/driver.go' files to restrict
the line length to 120 chars.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
Added helper func isNotMountPoint to check mountPoint,
validate error and reduce complexity of NodeStageVolume.
Signed-off-by: Rakshith R <rar@redhat.com>
This commit adds checks for missing `imageFeatures` parameter
in createvolumerequest and nodestagerequest(only for static PVs).
Missing `imageFeatures` parameter is ignored in case of non-static
PVs to ensure backwards compatibility with older versions which
did not have `imageFeatures` as required parameter.
Signed-off-by: Rakshith R <rar@redhat.com>
The Go linter paralleltest checks that the t.Parallel
gets called for the test method and for the range of
test cases within the test.
Updates: #2025
Signed-off-by: Yati Padia <ypadia@redhat.com>
While cleaning up snapshots, not all object may exist after a partial
provisioning attempt. In case objects are missing, do not try to delete
them.
Fixes: #2192
Signed-off-by: Niels de Vos <ndevos@redhat.com>
When cloning a volume from a (CSI) snapshot, we use DeepCopy() and do
not need an RBD snapshot as source.
Suggested-by: Madhu Rajanna <madhupr007@gmail.com>
Signed-off-by: Niels de Vos <ndevos@redhat.com>
In case restoring a snapshot of a thick-PVC failed during DeepCopy(),
the image will exist, but have partial contents. Only when the image has
the thick-provisioned metadata set, it has completed DeepCopy().
When the metadata is missing, the image is deleted, and an error is
returned to the caller. Kubernetes will automatically retry provisioning
on the ABORTED error, and the restoring will get restarted from the
beginning.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
the parent volume(CreateVolume) and the clone volume
(CreateSnapshot) are both indepedent and parent volume
can be deleted anytime. To check the thick provision
during Snapshot restore(CreateVolume from snapshot)
we need the thick provision metadata so for the same
reason setting the thick provision metadata on the
clone image we are creating at the CreateSnapshot time.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
RbdSnapName holds the actual RBD image name which
got created during the CreateSnapshot operation.
RbdImageName holds the name of the parent from
which the snapshot is created. and the parent
is independent of snapshot and it can be deleted
any time for the same reason using the RbdSnapName
to check the rbd image details.
generate a temporary volume from the snapshot which
replaces the rbdImageName with RbdSnapName and use
it to check the image metadata.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
added validation to allow only Restore of Thick PVC
snapshot to a thick clone and creation of thick clone
from thick PVC.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
isThickProvisioned can be used for both snapshot
and clone validation if isThickProvisioned is method
of common rbdImage structure.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
In case cloning a thick-PVC failed during DeepCopy(), the image will
exist, but have partial contents. Only when the image has the
thick-provisioned metadata set, it has completed DeepCopy().
When the metadata is missing, the image is deleted, and an error is
returned to the caller. Kubernetes will automatically retry provisioning
on the ABORTED error, and the cloning will get restarted from the
beginning.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Not all Linux kernels support the deep-flatten feature. Disabling the
feature makes it possible to map RBD images on older kernels (like what
minikube uses).
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Thick-provisioned images are independent, cloned images or snapshots are
deep-flattened during creation. There is no need to try and flatten them
again.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
To create a full-allocated RBD image from a snapshot/clone DeepCopy()
can be used. This is needed when the parent of the new volume is
thick-provisioner, so that the new volume is independent of the parent
and thick-provisioned as well.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
isCompatibleEncryption is used to validate the
requested volume and the existing volume and
the destination volume name wont be generated yet
and logging the destination volume prints the empty
image name with pool name.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
actual error will be present in the stdErr not the error
when we try to add a task to flatten the rbd image. This
commits corrects the error checking when the image does
not have a parent.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
This commit addresses the following issue:
'nolint:gocyclo // complexity needs to be reduced.'
is unused for linter "gocyclo" (nolintlint)
Updates:#2025
Signed-off-by: Yati Padia <ypadia@redhat.com>
While traversing image chain, the parent
image can be present in a different pool
that the one child is in. So, updating
pool name in the next itteration to
that of the Parent.
Co-authored-by: Madhu Rajanna <madhupr007@gmail.com>
Signed-off-by: Yug <yuggupta27@gmail.com>
Validate Snapshot request to check if the
passed pool name is not empty.
Co-authored-by: Madhu Rajanna <madhupr007@gmail.com>
Signed-off-by: Yug <yuggupta27@gmail.com>
if the parent and child clones are in
different namespaces we need to open a new
ioctx for pools.
Co-authored-by: Madhu Rajanna <madhupr007@gmail.com>
Signed-off-by: Yug <yuggupta27@gmail.com>
when clones are created in different pool,we
need to retrieve the parent pool to get the
information of the parent image.
Co-authored-by: Madhu Rajanna <madhupr007@gmail.com>
Signed-off-by: Yug <yuggupta27@gmail.com>
as we are supporting the creation of clone to a new
pool we need to pass the correct parent volume
to cleanup the snapshot on parent volume.
Co-authored-by: Madhu Rajanna <madhupr007@gmail.com>
Signed-off-by: Yug <yuggupta27@gmail.com>
added support to create image in different pool.
if the snapshot/rbd image exists in one pool we
can create a clone the clone of the rbd image to
a different pool.
Co-authored-by: Madhu Rajanna <madhupr007@gmail.com>
Signed-off-by: Yug <yuggupta27@gmail.com>
In golang world, if you split an empty string that does not contain
the seperator, you get an array with one empty string. This results
in volumes failing to mount with "invalid feature " (note extra space
because it's trying to check if 'empty string' is a valid feature).
This patch checks if the string is empty, and if so, it just decides
to skip the entire validation and returning nothing.
Signed-off-by: Mohammed Naser <mnaser@vexxhost.com>
CreateVolume will fail in below cases
* If the snapshot is encrypted and requested volume
is not encrypted
* If the snapshot is not encrypted and requested
volume is encrypted
* If the parent volume is encrypted and requested volume
is not encrypted
* If the parent volume is not encrypted and requested
volume is encrypted
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Move the repairing of a volume/snapshot from CreateVolume to its own
function. This reduces the complexity of the code, and makes the
procedure easier to understand. Further enhancements to repairing an
exsiting volume can be done in the new function.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
This commit addresses ifshort linter issues which
checks if short syntax for if-statements is possible.
updates: #1586
Signed-off-by: Rakshith R <rar@redhat.com>
This commit removes calling of .String() when logging
since `%s`,`%v` or `%q` will call an existing .String() function
automatically.
Fixes: #2051
Signed-off-by: Rakshith R <rar@redhat.com>
The NodeGetVolumeStats procedure can now be used to fetch the capacity
of the RBD block-device. By default this is a thin-provisioned device,
which means that the capacity is not reserved in the Ceph cluster. This
makes it possible to over-provision the cluster.
In order to detect the amount of storage used by the RBD block-device
(when thin-provisioned), it is required to connect to the Ceph cluster.
Unfortunately, the NodeGetVolumeStats CSI procedure does not provide
enough parameters to connect to the Ceph cluster and fetch more details
about the RBD image.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
flatten the image if the deep-flatten feature
is present on the images in the chain or if the
images in chain is not zero, as we cannot check
the deep-flatten feature the images which are
in trash.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
For flatten we call checkImageChainHasFeature
which internally calls to getImageInfo returns
the parent name even if the parent is in the trash,
when we try to open the parent image to get its
information it fails as the image not found.
we should treat error as nil if the parent is not found.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
The new gosec 2.7.0 complains like:
G304 (CWE-22): Potential file inclusion via variable (Confidence: HIGH, Severity: MEDIUM)
Updates: #2025
Signed-off-by: Niels de Vos <ndevos@redhat.com>
To recover from split brain (up+error) state the image need to be
demoted and requested for resync on site-a and then the image on site-b
should gets demoted.The volume should be marked to ready=true when the
image state on both the clusters are up+unknown because during the last
snapshot syncing the data gets copied first and then image state on the
site-a changes to up+unknown.
If the image state on both the sites are up+unknown consider that
complete data is synced as the last snapshot
gets exchanged between the clusters.
* create 10 GB of file and validate the data after resync
* Do Failover when the site-a goes down
* Force promote the image and write data in GiB
* Once the site-a comes back, Demote the image and issue resync
* Demote the image on site-b
* The status will get reflected on the other site when the last
snapshot sync happens
* The image will go to up+unknown state. and complete data will
be copied to site a
* Promote the image on site-a and use it
```bash
csi-vol-5633715e-a7eb-11eb-bebb-0242ac110006:
global_id: e7f9ec55-06ab-46cb-a1ae-784be75ed96d
state: up+unknown
description: remote image demoted
service: a on minicluster1
last_update: 2021-04-28 07:11:56
peer_sites:
name: e47e29f4-96e8-44ed-b6c6-edf15c5a91d6-rook-ceph
state: up+unknown
description: remote image demoted
last_update: 2021-04-28 07:11:41
```
* Do Failover when the site-a goes down
* Force promote the image on site-b and write data in GiB
* Demote the image on site-b
* Once the site-a comes back, Demote the image on site-a
* The images on the both site will go to split brain state
```bash
csi-vol-37effcb5-a7f1-11eb-bebb-0242ac110006:
global_id: 115c3df9-3d4f-4c04-93a7-531b82155ddf
state: up+error
description: split-brain
service: a on minicluster2
last_update: 2021-04-28 07:25:41
peer_sites:
name: abbda0f0-0117-4425-8cb2-deb4c853da47-rook-ceph
state: up+error
description: split-brain
last_update: 2021-04-28 07:25:26
```
* Issue resync
* The images cannot be resynced because when we issue resync
on site a the image on site-b was in demoted state
* To recover from this state (promote and then demote the
image on site-b after sometime)
```bash
csi-vol-37effcb5-a7f1-11eb-bebb-0242ac110006:
global_id: 115c3df9-3d4f-4c04-93a7-531b82155ddf
state: up+unknown
description: remote image demoted
service: a on minicluster1
last_update: 2021-04-28 07:32:56
peer_sites:
name: e47e29f4-96e8-44ed-b6c6-edf15c5a91d6-rook-ceph
state: up+unknown
description: remote image demoted
last_update: 2021-04-28 07:32:41
```
* Once the data is copied we can see that the image state
is moved to up+unknown on both sites
* Promote the image on site-a and use it
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
when a Snapshot is encrypted during a CreateSnapshot
operation, the encryption key gets created in the KMS
when we delete the Snapshot the key from the KMS
should also gets deleted.
When we create a volume from snapshot we are copying
required information but we missed to copy the
encryption information, This commit adds the missing
information to delete the encryption key.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
At present we return the volume connect error if the clone
from snapshot fails when rbdvolume is encrypted, which is incorrect.
This patch correctly return the failed copy encryption error to the
caller
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
Before RBD map operation, we do check the
watchers on the RBD image. In the case of
RWO volume. cephcsi makes sure only one
client is using the RBD image. If the rbd
image is mirrored, by default mirroring
daemon will add a watcher on the image
and as we are using go-ceph a watcher will
be added as we have opened the image So
we will have two watchers on an image if
mirroring is enabled. This holds when the
rbd mirror daemon is running, In case if
the mirror daemon is not running there will
be only one watcher on the rbd image
(which is placed by go-ceph image open)
we should not block the map operation if
the mirroring daemon is not running as
its Async mirroring. This commit adds a
check to make sure no more than 2 watchers
if the image is mirrored or no more than 1
watcher if it is not mirrored image.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Since rbdImage is a common struct for
rbdVolume and rbdSnapshot, it description
was matching to only snapshot.
This commit makes the comments generic for
both volumes and snapshots.
Signed-off-by: Yug <yuggupta27@gmail.com>
If the pool or few keys are missing in the omap.
GetImageAttributes function returns nil error message and few
empty items in imageAttributes struct. if the image is not
found and the entiries are missing use
the volumeId present on the PV annotation for further operations.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
incase if the image is promoted and demoted the
image state will be set to up+unknown if the image
on the remote cluster is still in demoted state.
when user changes the state from primary to secondary
and still the image is in demoted (secondary) state
in the remote cluster. the image state on both the cluster
will be on unknown state.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
It helps to get a stack trace when debugging issues. Certain things are
considered bugs in the code (like missing attributes in a struct), and
might cause a panic in certain occasions.
In this case, a missing string will not panic, but the behaviour will
also not be correct (DEKs getting encrypted, but unable to decrypt).
Clearly logging this as a BUG is probably better than calling panic().
Signed-off-by: Niels de Vos <ndevos@redhat.com>
It is possible that when a provisioner restarts after a snapshot was
cloned, but before the newly restored image had its encryption metadata
set, the new image is not marked as encrypted. This will prevent
attaching/mounting the image, as the encryption key will not be fetched,
or is not available in the DEKStore.
By actively repairing the encryption configuration when needed, this
problem should be addressed.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
buildCreateVolumeResponse() exists exactly for the need to create a
csi.CreateVolumeResponse based on an rbdVolume. Calling this helper
reduces the code duplication in CreateVolume().
Signed-off-by: Niels de Vos <ndevos@redhat.com>
The rbdVolume that needs its encryption configured is constructed in the
Exists() method. It is suitable to move the copyEncryptionConfig() call
there as well, so that the object is completely constructed in a single
place.
Golang-ci:gocyclo complained about the increased complexity of the
Exists() function. Moving the repairing of the ImageID into its own
helper function makes the code a little easier to understand.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Introduce helper function cloneFromSnapshot() that takes care of the
procedures that are needed when an existing snapshot has been found.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
When a source volume is encrypted, the passphrase needs to be copied and
stored for the newly cloned volume.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Cloning volumes requires copying the DEK from the source to the newly
cloned volume. Introduce copyEncryptionConfig() as a helper for that.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Without this, the rbdVolume can not connect to the Ceph cluster and
configure the (optional) encryption.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
The ControllerServer should not need to care about support for
encryption, ideally it is transparantly handled by the rbdVolume type
and its internal API.
Deleting the DEK was one of the last remainders that was explicitly done
inside the ControllerServer.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
In the case of the Async DR, the volumeID will
not be the same if the clusterID or the PoolID
is different, With Earlier implementation, it
is expected that the new volumeID mapping is
stored in the rados omap pool. In the case of the
ControllerExpand or the DeleteVolume Request,
the only volumeID will be sent it's not possible
to find the corresponding poolID in the new cluster.
With This Change, it works as below
The csi-rbdplugin-controller will watch for the PV
objects, when there are any PV objects created it
will check the omap already exists, If the omap doesn't
exist it will generate the new volumeID and it checks for
the volumeID mapping entry in the PV annotation, if the
mapping does not exist, it will add the new entry
to the PV annotation.
The cephcsi will check for the PV annotations if the
omap does not exist if the mapping exists in the PV
annotation, it will use the new volumeID for further
operations.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
It seems that newer versions of some tools/libraries identify encrypted
filesystems with `crypto_LUKS` instead of `crypt`.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
In case of the DR the image on the primary site cannot be
demoted as the cluster is down, during failover the image need
to be force promoted. RBD returns `Device or resource busy`
error message if the image cannot be promoted for above reason.
Return FailedPrecondition so that replication operator can send
request to force promote the image.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
instead of fetching the force option from the
parameters. Use the Force field available in
the PromoteVolume Request.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
as the org github.com/kube-storage is renamed
to github.com/csi-addons as the name kube-storage
was more generic.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Incase of resync the image will get deleted, gets
recreated and its a a time consuming operation.
It makes sense to return aborted error instead
of not found as we have omap data only the image
is missing in rbd pool.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Do resync if the image is in unknow or in error
state.
Check for the current image state for up+stopped
or up+replaying and also all peer site status
should be un up+stopped to confirm that resyncing
is done and image can be promoted and used.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
added replication related operations as a method
of rbdImage as these methods can be easily used
when we introduce volumesnaphot mirroring operations.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
the rbd mirror state can be in enabled,disabled
or disabling state. If the mirroring is not disabled
yet and still in disabling state. we need to
check for it and return abort error message if
the mirroring is still getting disabled.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
added ReplicationServer struct for the replication related
operation it also embed the ControllerServer which
already implements the helper functions like locking/unlocking etc.
removed getVolumeFromID and cleanup functions for better
code readability and easy maintaince.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
There is no need for each EncryptionKMS to implement the same GetID()
function. We have a VolumeEncryption type that is more suitable for
keeping track of the KMS-ID that was used to get the configuration of
the KMS.
This does not change any metadata that is stored anywhere.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Current rbd plugin only supports the layering feature
for rbd image. Add exclusive-lock and journaling image
features for the rbd.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Signed-off-by: woohhan <woohyung_han@tmax.co.kr>
rbdVolumes can have several resources that get allocated during its
usage. Only destroying the IOContext may not be suffiecient and can
cause resource leaks.
Use rbdVolume.Destroy() when the rbdVolume is not used anymore.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Connections are reference counted, so just assigning the connection to
an other object for re-use is not correct. This can cause connections to
be garbage collected while something else is still using it.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
as RBD is implementing the replication
we are registering it. For CephFS, its
not implementing the replication we are
passing nil so we dont want to register
it.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Because rbdVolume and rbdSnapshot are very similar, they can be based
off a common struct rbdImage that contains the common attributes and
functions.
This makes it possible to re-use functions for snapshots, and prevents
further duplication or code.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
The rbdSnapshot and rbdVolume structs have many common attributes. In
order to combine these into an rbdImage struct that implements shared
functionality, having the same attribute for the ID makes things much
easier.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
NewVolumeEncryption() will return an indication that an alternative
DEKStore needs to be configured in case the KMS does not support it.
setKMS() will also set the DEKStore if needed, so renaming it to
configureEncryption() makes things clearer.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Use DEKStore API for Fetching and Storing passphrases.
Drop the fallback for the old KMS interface that is now provided as
DEKStore. The original implementation has been re-used for the DEKStore
interface.
This also moves GetCryptoPassphrase/StoreNewCryptoPassphrase functions
to methods of VolumeEncryption.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Prepare for grouping encryption related functions together. The main
rbdVolume object should not be cluttered with KMS or DEK procedures.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
For NodeUnstageVolume its a two step process,
first unmount the volume and than unmap the volume.
Currently, we are logging only after rbd unmapping is done.
sometimes it becomes difficult to debug with above logging
whether more time is spent in unmount or unmap.
This commits adds one more debug log after unmount is done.
with this we can identify where exactly more time is spent
by looking at the logs.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
It seems that writing more than 1 GiB per WriteSame() operation causes
an EINVAL (22) "Invalid argument" error. Splitting the writes in blocks
of maximum 1 GiB should prevent that from happening.
Not all volumes are of a size that is the multiple of the stripe-size.
WriteSame() needs to write full blocks of data, so in case there is a
small left-over, it will be filled with WriteAt().
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Introduce initKMS() as a function of rbdVolume. KMS functionality does
not need to pollute general RBD image functions. Encryption functions
are now in internal/rbd.encryption.go, so move initKMS() there as well.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
When and RBD image is expanded, the additional extents need to get
allocated when the image was thick-provisioned.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
When images get resized/expanded, the additional space needs to be
allocated if the image was initially thick-provisioned. By marking the
image with a "thick-provisioned" key in the metadata, future operations
can check the need.
A missing "thick-provisioned" key indicates that the image has not been
thick-provisioned.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Write blocks of stripe-size to allocate RBD images when
Thick-Provisioning is enabled in the StorageClass.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Add an option to the StorageClass to support creating fully allocated
(thick provisioned) RBD images
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Signed-off-by: Niels de Vos <ndevos@redhat.com>
When a volume was provisioned by an old Ceph-CSI provisioner, the
metadata of the RBD image will contain `requiresEncryption` to indicate
a passphrase needs to be created. New Ceph-CSI provisioners create the
passphrase in the CreateVolume request, and set `encryptionPrepared`
instead.
When a new node-plugin detects that `requiresEncryption` is set in the
RBD image metadata, it will fallback to the old behaviour.
In case `encryptionPrepared` is read from the RBD image metadata, the
passphrase is used to cryptsetup/format the image.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
This adds internal/rbd/encryption.go which will be used to include other
encryption functionality to support additional KMS related functions. It
will work together with the shared API from internal/util/kms.go.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Have the provisioner create the passphrase for the volume, instead of
doign it lazily at the time the volume is used for the 1st time. This
prevents potential races where pods on different nodes try to store
different passphrases at the (almost) same time.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
It seems that calls to addRbdManagerTask() do not include the
rados-namespace in the image location. Functions calling
addRbdManagerTask() construct the image location themselves, but should
use rbdVolume.String() to include all the attributes.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Make rbdplugin pod work in a non-initial network namespace (i.e. with
"hostNetwork: false") by skipping waiting for udev events when mapping
and unmapping images. CSI use case is very simple: all that is needed
is a device node which is immediately fed to mkfs, so we should be able
to tolerate udev not being finished with the device just fine.
Fixes: #1323
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
With the new support for passing --options, referring to ExecCommand()
argument slices as mapOptions and options is confusing.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
mount packges is moved from
k8s.io/utils/mount to a new repository
k8s.io/mount-utils. updated code to use
the same.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Add a new method to the EncryptionKMS interface so that resources can be
freed when EncryptionKMS instances get freed.
With the move to using the libopenstorage API, a temporary file needs to
store the optional CA certificate. The Destroy() method of the
vaultConnection type now removes this file.
The rbdVolume uses the EncryptionKMS type now, so call the new Destroy()
method from withing rbdVolume.Destroy().
Signed-off-by: Niels de Vos <ndevos@redhat.com>
The Owner of an RBD image (Kubernetes Namespace, tenant) can be used to
identify additional configuration options. This will be used for
fetching the right Vault Token when encrypting/decrypting volumes.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
If cephcsi encounters any error after
reservation, as a cleanup operation
it should revert back the reservation.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
If the omap data already exits return nil.
so that omap generator will not try to reserve
anything again.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
earlier if the depth check fails the
complete vol struct was getting logged,
this commits logs only the pool and image
name.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
In the case of Disaster Recovery failover, the
user expected to create the static PVC's. We have
planned not to go with the PVC name and namespace
for many reasons (as in kubernetes it's planned to
support PVC transfer to a new namespace with a
different name and with new features coming in
like data populator etc). For now, we are
planning to go with static PVC's to support
async mirroring.
During Async mirroring only the RBD images are
mirrored to the secondary site, and when the
user creates the static PVC's on the failover
we need to regenerate the omap data. The
volumeHandler in PV spec is an encoded string
which contains clusterID and poolID and image UUID,
The clusterID and poolID won't remain same on both
the clusters, for that cephcsi need to generate the
new volume handler and its to create a mapping
between new volume handler and old volume handler
with that whenever cephcsi gets csi requests it
check if the mapping exists it will pull the new
volume handler and continues other operations.
The new controller watches for the PVs created,
It checks if the omap exists if it doesn't it
will regenerate the entire omap data.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
in case of mirrored image, if the image is
primary a watcher will be added by the rbd
mirror deamon on the rbd image.
we have to consider 2 watcher to check image
is in use.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
incase of async mirroring the volume UUID is
retrieved from the volume name, instead of cephcsi
generating a new UUID it should reserve the passed
UUID it will be useful when we support both metro DR
and async mirroring on a kubernetes clusters.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
An rbd image can have a maximum number of
snapshots defined by maxsnapshotsonimage
On the limit is reached the cephcsi will
start flattening the older snapshots and
returns the ABORT error message, The Request
comes after this as to wait till all the
images are flattened (this will increase the
PVC creation time. Instead of waiting till
the maximum snapshots on an RBD image, we can
have a soft limit, once the limit reached
cephcsi will start flattening the task to
break the chain. With this PVC creation time
will only be affected when the hard limit
(minsnapshotsonimage) reached.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
`cap` builtin function returns the capacity of a type. Its not
good practice to use this builtin function for other variable
names, removing it here
Ref# https://golang.org/pkg/builtin/#cap
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
volJournal.Connect() got the error on err2 variable, however
the return was on variable err which hold the error return of
DecomposeCSIID() which is wrong. This cause the error return wrongly
parsed and pushed from the caller. From now on, we are reusing the
err variable to hold and revert the error of volJournal.Connect().
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
All the previous condition checks exit from the function and
when it reach to this block its obvious that error is non nil,
we dont need an extra check here.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
The error check condition in genVolFromID() is always false as far as
code reading/workflow goes. Removing it for the same reason.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
cephcsi uses cli for fetching snap list as well as to check the
snapshot namespace, replaced that with go-ceph calls.
Signed-off-by: Mudit Agarwal <muagarwa@redhat.com>
It doesnot make sense to allow the creation of empty
volumes which is going to be accessed with readonly mode,
this commit allows the creation of volume which is having
readonly capabilities only if the content source is set
for the volume.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Form kubernetes v1.19 onwards NodeRequest is getting volume path
in StagingTargetPath instead of VolumePath, cephcsi should also
use the same.
Signed-off-by: Mudit Agarwal <muagarwa@redhat.com>
updateVolWithImageInfo() is currently executing an rbd command
to get details about an RBD image. Replaced it with the
required go-ceph functions.
Signed-off-by: Mudit Agarwal <muagarwa@redhat.com>
in case of clone failure, we need to first delete
the clone and the snapshot from which we created
the clone, then as part of cleanup we need to remove
the temporary cloned image and the temporary snapshot
created on the parent image.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
At CSI spec < 1.2.0, there was no volumecapability in the
expand request. However its available from v1.2+ which allows
us to declare the node operations based on the volume mode.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
if the image is created without flattening image-feature
the image will get few image-features by default, deep-flatten
is one of them. if the image doesnot have any parent
the rbd image flattening will fail, This commit discards
error message if the image doesnt have any parent.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
merging of https://github.com/ceph/ceph-csi/pull/1035
broken the cephcsi building. This commits fixes
the build issue.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Make sure to operate within the namespace if any given
when dealing with rbd images and snapshots and their journals.
Signed-off-by: Mehdy Khoshnoody <mehdy.khoshnoody@gmail.com>
if we are not able to fetch the cluster-ID from
the createSnapshot request and also if we are
not able to get the monitor information from
the cluster-ID return error instead of using
the parent image information.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Replaced command execution with go-ceph Resize() function.
Volsize is being updated before waiting for resize() to return,
fixed it to get updated only after resize() is successful.
Signed-off-by: Mudit Agarwal <muagarwa@redhat.com>
While adding the context.Context to the resizeRBDimage() function, it
became a little ugly. So renaming the function to resize() and making it
a method of the rbdVolume type.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Most consumers of util.ExecCommand() need to convert the returned []byte
format of stdout and/or stderr to string. By having util.ExecCommand()
return strings instead, the code gets a little simpler.
A few commands return JSON that needs to be parsed. These commands will
be replaced by go-ceph implementations later on. For now, convert the
strings back to []byte when needed.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
All calls to util.ExecCommand() now pass the context.Context. In some
cases this is not possible or needed, and util.ExecCommand() will not
log the command.
This should make debugging easier when command executions fail.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Current versions of the mkfs.xfs binary enable reflink support by
default. This causes problems on systems where the kernel does not
support this feature. When the kernel the feature does not support, but
the filesystem has it enabled, the following error is logged in `dmesg`:
XFS: Superblock has unknown read-only compatible features (0x4) enabled
Introduce a check to see if mkfs.xfs supports the `-m reflink=` option.
In case it does, pass `-m reflink=0` while creating the filesystem.
The check is executed once during the first XFS filesystem creation. The
result of the check is cached until the nodeserver restarts.
Fixes: #966
Signed-off-by: Niels de Vos <ndevos@redhat.com>
The new rbdVolume.isInUse() method will replace the rbdStatus()
function. This removes one more rbd command execution in the
DeleteVolume path.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
This change replaces the sentinel errors in rbd module with
standard errors created with errors.New().
Related: #1203
Signed-off-by: Sven Anderson <sven@redhat.com>
The sentinel error code had additional fields in the errors, that are
used nowhere. This leads to unneccesarily complicated code. This
change replaces the sentinel errors in utils with standard errors
created with errors.New() and adds a simple JoinErrors() function to
be able to combine sentinel errors from different code tiers.
Related: #1203
Signed-off-by: Sven Anderson <sven@redhat.com>
Take operation locks on the resources before operating
on the resouces. This allows us to do parallel operations
for some RPC calls such as Clone and Restore of PVC.
This operations will only be blocked if the image is
expanding or Snapshot and RBD image is getting deleted.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Direct usage of numbers should be avoided.
Issue reported:
mnd: Magic number: X, in <argument> detected (gomnd)
Signed-off-by: Yug <yuggupta27@gmail.com>
