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>
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>
The implementation of getOMapValues assumed that the number of key-value
pairs assigned to the object would be close to the number of keys
being requested. When the number of keys on the object exceeded the
"listExcess" value the function would fail to read additional keys
even if they existed in the omap.
This change sets a large fixed "chunk size" value and keeps reading
key-value pairs as long as the callback gets called and increments
the numKeys counter.
Signed-off-by: John Mulligan <jmulligan@redhat.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>
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>
Direct usage of numbers should be avoided.
Issue reported:
mnd: Magic number: X, in <argument> detected (gomnd)
Signed-off-by: Yug <yuggupta27@gmail.com>
Several places in the code compared errors directly with the go-ceph
sentinel errors. This change uses the errors.Is() function of go
1.13 instead. The err113 linter reported this issue as:
err113: do not compare errors directly, use errors.Is() instead
Signed-off-by: Sven Anderson <sven@redhat.com>
"github.com/pkg/errors" does not offer more functionlity than that we
need from the standard "errors" package. With Golang v1.13 errors can be
wrapped with `fmt.Errorf("... %w", err)`. `errors.Is()` and
`errors.As()` are available as well.
See-also: https://tip.golang.org/doc/go1.13#error_wrapping
Signed-off-by: Niels de Vos <ndevos@redhat.com>
cephcsi need to store and retrieve the rbd image ID
in the omap as we need the image ID to add a
task to remove the image from the Trash.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
The previous function used to remove omap keys apparently did not
return errors when removing omap keys from a missing omap (oid).
Mimic that behavior when using the api.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
For any function that sets more than one key on a single oid setting
them as a batch will be more efficient.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
For any function that removes more than one key on a single oid removing
them as a batch will be more efficient.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
Taking this appraoch means that any function that must get more than one
key's value from the same oid can be more efficient by calling out to
ceph only once.
To be cautious and avoid missing things we always request ceph return
more keys than we actually expect to be set on the oid. If there are
unexpected keys there, we will not miss the keys we want if we first hit
an unexpected key if we were to limit ourselves to iterating only over
the number of keys we're expecting to be on the object.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
Convert the business-logic of the journal to use the new go-ceph based
omap manipulation functions.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
These new omap manipulation functions (get/set/remove) are roughly
equivalent to the previous command-line based approach but rely
on direct api calls to ceph.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
Before, the one CSIJournal type was handling both configuration and
providing methods to make changes to the journal. This created the
temptation to modify the state of the global configuration object to
enact changes through the method calls.
This change creates a new type `journal.Connection` that takes the
monitors and credentials to create a short(er)-lived object to actually
read and make changes on the journal. This also avoid mixing the
arguments needed to connect to the cluster with the arguments needed
for the various journal read & update calls.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
The SetNamespace setter function was called only once, immediately after
the creation of a volume journal object in cephfs only.
Remove this function so that it is no longer implied that this field can
be mutated after the journal is created. In it's place, use an extended
"constructor" NewCSIVolumeJournalWithNamespace that takes a namespace
value at create-time only.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
The function SetCSIDirectorySuffix was used only one per (long-lived,
gloabl) journal object. It is simpler to construct the journal objects
with this needed parameter:
1. As it is required to function and non-optional AFAICT
2. Removes the temptation to mutate global object
3. Reduces LOC with exact same functionality
4. SetCSIDirectorySuffix would not behave correctly if called a 2nd time
anyway.
Point 4. means that if you called the function twice to change the
suffix when you previously had "csi.volumes.alice", you'd get
"csi.volumes.alice.bob" instead of "csi.volumes.bob" what one would
expect.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
This new journal package isolates journal logic from the rest of util
and helps draw bright lines between what is a generic utility function
and what is csi journal logic.
Done partly as preparation for making use of go-ceph in journal.
No functional changes are made except to update references to allow the
code to compile.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
