assignMons schedules each mon in its own goroutine and shares a
failedMonSchedule bool to record whether any of them failed. The
resultLock mutex already guards the shared c.mapping.Schedule map
write, but the three failedMonSchedule = true assignments were done
without holding the lock. When two or more mons fail to schedule at
the same time (for example when waitForMonitorScheduling errors, the
node choice is nil, or getNodeInfoFromNode fails) their goroutines
write the flag concurrently, which is a write-write data race.
Guard the failedMonSchedule writes with the existing resultLock, the
same lock the goroutines already use for the map update. The post-Wait
read is left as is since the WaitGroup orders it after every write.
Add a regression test that fails scheduling for multiple mons at once
and confirms assignMons returns an error; run it with -race to catch
the race.
Signed-off-by: Anas Khan <83116240+anxkhn@users.noreply.github.com>
Several godoc comments led with a stale or incorrect identifier, left
over from renames, exported/unexported changes, copy-paste between
sibling declarations, or plain typos. As a result the documented name no
longer matched the function, method, type, or var it describes. Correct
each leading word to the name of the declaration it documents.
Signed-off-by: Joshua Hoblitt <josh@hoblitt.com>
Going forward, admin will manage the csi operator
CR's and rook will only manage Ceph Connection cr
and client Profile cr.
The old csi driver is completely removed from Rook
and can no longer be used starting in Rook v1.20.
The upgrade guide will contain the needed transition steps
for managing the csi operator settings.
Signed-off-by: subhamkrai <srai@redhat.com>
Ceph relaxed the stretch mode CRUSH weight check in ceph/ceph#66580 to
allow weights within 10 percent of each other. This change is available
in Ceph Umbrella and was backported to tentacle v20.2.2 via ceph/ceph
number 67790. The squid backport is still pending.
Signed-off-by: Lumir Sliva <61183145+lumir-sliva@users.noreply.github.com>
The mon PDB was previously only being adjusted
during mon failover, where the maxunavailable was set to 0
before failover, then reset to 1 after failover. This allowed
for a race condition where one mon had been drained and is
not fully back in quorum before the next mon is allowed to drain.
This is now prevented by disallowing drains until a mon
is fully back in quorum. Anytime a mon falls out of quorum,
the mon pdb will be adjusted to prevent further drains
from causing a mon quorum outage.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
for tnf, adding initial api changes and implementation for floating mon.
The current pr follow embeded yaml approach for creating floating mon
deploymen instead of configmap approach.
Signed-off-by: subhamkrai <srai@redhat.com>
The do-not-reconcile label on the mon deployments was
causing the mon reconcile to be skipped for all mons,
even if the label was only on a single mon. Now, the
skipping of reconcile will only be scoped to a single
mon.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
When writing to the operator log, where possible
the namespace of the resource on which the operator
is working will be logged as a standard prefix
to more easily troubleshoot what resource is being reconciled
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
Wait for the mon-canary pods to be terminated before starting the actual
mons. When topology spread constraints are added to evenly spread the
mons, then pod-anti affinity does not allow mons to be created on the
same node where other mons are running, this also includes the
mon-canary pods. We delete the mon-canary pods and then start the actual
mon pods. Sometimes the termination of the canary pods might take more
time, so schedular won't schedular mon pods on that node.
Signed-off-by: Santosh Pillai <sapillai@redhat.com>
The mon canary should quickly shut down after the pod is
deleted, since it has no purpose to continue running.
Since the sleep does not catch the SIGTERM as pid 1,
use a bash trap to exit quickly.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
Allow Rook to rotate CephX keys for Ceph v19.2.3 and up with the
exception of Ceph mons. While v19.2.3 has the `ceph auth rotate` command
that Rook needs to perform key rotation, there is still ongoing work in
Ceph to change how the `mon.` key is tracked internally. Rook's mon key
rotation requires those currently-unreleased Ceph changes.
Signed-off-by: Blaine Gardner <blaine.gardner@ibm.com>
Cleanup up overall usage of CephX statuses in CephCluster.
Convert all CephX statuses to non-pointer types. There is no specific
need for pointer types, and keeping these as non-pointers means there is
not a risk of nil pointer exceptions in all the various Rook controllers
for upgraded clusters.
Move initialization of CephCluster CephX status items to
`preMonStartupActions()`. This helps resolve 2 issues:
1. 'Uninitialized' state was being set for external CephClusters where
CephX statuses are not relevant.
2. 'Uninitialized' state was being set for upgraded clusters whose key
status cannot be determined.
Signed-off-by: Blaine Gardner <blaine.gardner@ibm.com>
CephConnection CR is created with the
kubernetes namespace name as its the
unique name in the entire cluster.
This PR removes unwanted input
parameters and adjust the namespace
name for the cephConnection CR.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
When a mon is failed over, the old mon is left in mon quorum
until the new quorum is confirmed in quorum. When host networking
is enabled, this may result in two mons being in quorum with the
same ip, if the new mon is scheduled on the same node as the
down mon. This results in daemon crashes in the cluster since
they are confused by multiple of the same ip in the mon map.
The solution is to avoid scaling down the bad mon when host networking
is enabled. This will ensure the new mon is not scheduled on the
same node as the previous mon. With host networking, either another
node is necessary or the original mon must come back online to
restore quorum.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
This is the first implementation of CephX key rotation in Rook.
Adds key rotation API from design PR 15915.
Also add helper methods for rotating keys, determining when keys
need to be rotated, and for generating CephX key statuses. These will be
reusable for other Rook reconciles beyond object/RGW.
Signed-off-by: Blaine Gardner <blaine.gardner@ibm.com>
This change cleans up and standardizes import statements across the Ceph operator code. It removes redundant or duplicate imports and reorganizes alias names for improved clarity and consistency. Additionally, the ST1019 exception was removed from .golangci.yaml now that the code complies with the rule.
Signed-off-by: Carlos Barria <cbarria@yahoo.com>
With a osd resize operation, currently pvc and pv may
takes more time then expected.
And so the osd dont get resized and need manual re-start
as currently we don't have a watch on osd's
controller that watches pv and get it resized.
add a wait time till pvc is not resized
Signed-off-by: parth-gr <partharora1010@gmail.com>
The change creates and manages an Endpoints resource
with the current Ceph monitor (mon) IPs, allowing
clients to resolve mon IPs via DNS without relying
on the rook-ceph-mon-endpoints ConfigMap.
Signed-off-by: Patryk Rostkowski <patrostkowski@gmail.com>
Implements #14733. Allows to set IDs of external mons to
Cluster CRD. Rook will not remove external mons from quorum
and will add external mon addresses to mon endpoints.
Use-case for external mon is to maintain quorum for 2-AZ
k8s cluster in case of zone outage.
Signed-off-by: Artem Torubarov <artem.torubarov@clyso.com>
There is a race condition in ceph v19 at startup of a
stretch cluster where if the .mgr pool is created by
the mgr pod after stretch mode is entered, the
devicehealth mgr module will crash and cause a
ceph HEALTH_ERR. The error then causes rook reconcile
of all CRs to be skipped because Rook is assuming
admin intervention is needed to continue. The fix
is to wait until the .mgr pool is created by the mgr
pod before entering stretch mode.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
If the mon failover is in progress, ensure the removal
of an extra mon deployment is skipped since that code
path only has one mon in the list for the mon that was
just newly started. The extra mon was erroneously removing
a random mon in that case, followed immediately by the mon
failover completing and removing the expected failed mon,
and potentially causing mon quroum loss.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
When removing a mon from quorum, there is a race condition that
can result in mon quorum going being lost at least temporarily.
The mon pod was being deleted first, and then the mon removed
from quorum. If any other mon went down between the time the
pod of the bad mon was deleted and when the mon was removed from
quorum, there may not be sufficient quorum to complete the
action of removing the mon from quorum and the operator would
be stuck.
For example, there could be 4 mons temporarily due to timing
of upgrading K8s nodes where mons may be taken down for some
number of minutes. Say a new mon is started while the down
mon also comes back up. Now the operator sees it can remove
the 4th mon from quorum, so it starts to remove it. Now say
another mon goes down on another node that is being updated
or otherwise drained. Since the 4th mon pod was deleted
and another mon is down, there are only two mons remaining
in quorum, but 3 mons are required in quorum when there
are 4 mons. Therefore, the quorum is stuck until the
third mon comes back up.
The solution is to first remove the extra mon from quorum
before taking down the mon pod.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
adding code changes,rbac changes required for create the new
Ceph-CSI operator CR named cephCluster in api group 'csi.ceph.io'.
Signed-off-by: subhamkrai <srai@redhat.com>
Someone testing the Multus holder pod removal feature encountered an
issue where the migration process failed to lead to a system state where
PVCs could be created successfully.
The root cause was found to be a ceph csi config map wherein the primary
CephCluster entry was lacking a value for the "namespace" field.
I observed this once in my development on the holder pod removal
feature, but I was unable to reproduce and assumed it was my own error.
Since this has been seen in a user environment, it must be that the
error is a race condition, and I am unable to determine the exact source
of the bug.
I do not believe this bug would be present if the code that updates the
CSI configmap were properly idempotent, but it has many conditions based
on prior states, and I was unable to determine how to resolve this
underlying impelementation pattern issue.
Instead, I opted to separate the `clusterKey` parameter into two clear
parts:
1. `clusterID` for when `clusterKey` is used as an analogue for
`clusterID`
2. `clusterNamespace` for when `clusterKey` is used as an analogue for
`clusterNamespace`
I added unit tests to ensure that SaveClusterConfig() will be able to
detect when the namespace is currently missing, and using the new
`clusterNamespace` field, it should always know what value to use as
input when correcting the bug in already-installed systems.
I also verified that this update works when the function simultaneously
removes netNamespaceFilePath entries, and that those entries are removed
properly.
Finally, manual testing also verifies the change.
Signed-off-by: Blaine Gardner <blaine.gardner@ibm.com>
Reduce CRD size by removing some descriptions for things that
(a) are long, (b) are repeated, and (c) are commong kubernetes types.
Remove the `Status` field from VolumeClaimTemplates (a.k.a. PVCs) in
types.go to remove all unnecessary status-related fields.
Use `yq` to go through generated `crds.yaml` to strip out descriptions
for well-known k8s types that are long and repeated where we can't
remove them from types.go. This is mostly placement descriptions and
volumeSource descriptions.
The overall `crds.yaml` file is reduced to less than 60% of its original
size.
Signed-off-by: Blaine Gardner <blaine.gardner@ibm.com>
Signed-off-by: Redouane Kachach <rkachach@redhat.com>
Co-authored-by: Blaine Gardner <blaine.gardner@ibm.com>
This commit uses an csi config map API exposed by Ceph-CSI.
This will reduce the efforts of keeping the configuration in
sync with Ceph-CSI.
Signed-off-by: Praveen M <m.praveen@ibm.com>
During certain maintenance tasks the admin will own running
operations on the ceph mgr, rgw, mds and rbd-mirror daemons
and the operator should not interfere with those operations.
Co-authored-by: gauravsitlani <gaurav.sitlani@live.com>
Signed-off-by: subhamkrai <srai@redhat.com>
Change how Rook detects network CIDRs for Multus networks. The IPAM
configuration is only defined as an arbitrary string JSON blob with a
"type" field and nothing more. Rook's detection of CIDRs for whereabouts
had already grown out of date since the initial implementation.
Additionally, Rook did not support DHCP IPAM, which is a reasonable
choice for users. And more, Rook did not support CNI plugin chaining,
which further complicates NADs. Based on the CNI spec, network chaning
can result in any changes to network CIDRs from the first-given plugin.
All these problems make it more and more difficult for Rook to support
Multus by inspecting the NAD itself to predict network CIDRs. Instead,
it is better for Rook to treat the CNI process as a black box. To
preserve legacy functionality of auto-detecting networks and to make
that as robust as possible, change to a canary-style architecture like
that used for Ceph mons, from which Rook will detect the network CIDRs
if possible.
Also allow users to specify overrides for CIDR ranges. This allows Rook
to still support esoteric and unexpected NAD or network configurations
where a CIDR range is not detectable or where the range detected would
be incomplete. Because it may be impossible for Rook to understand the
network CIDRs wholistically while residing only on a portion of the
network, this feature should have been present from Multus's inception.
Improving CIDR auto-detection and allowing users to specify overrides
for auto-detected CIDRs rounds out Rook's Multus support for CephCluster
(core/RADOS) installations. No further architectural changes should be
needed for CephClusters as regards application of public/cluster network
CIDRs for Multus networks.
Signed-off-by: Blaine Gardner <blaine.gardner@ibm.com>
* fixes mon zonal deployment of 2 per zone to 1 if zones field is
specified.
* adds zones field doc to CRDs Documentation.
* add code for picking up failureDomainLabel if specified for zones
field in mon deployment
Signed-off-by: Deepika Upadhyay <deepika@koor.tech>