By setting the working directory of a Ceph daemon to the log direcor
(which is bindmounted to the host), we can ensure that the coredumps
will be available.
On CentOS 7 **only**, the kernel is configured with:
```
cat /proc/sys/kernel/core_pattern
core
```
which means that the coredumps will end up in the process working
directory and will be named "core".
On CentOS 8, everything is different and this won't work but still
CentOS will be able to consume this and changing the working directory
is harmless.
Signed-off-by: Sébastien Han <seb@redhat.com>
The arbiter can only be configured with the stretch cluster if the
CRUSH map is balanced and there are two zones in the CRUSH map.
After the OSDs are configured, we wait for all the OSD pods to be
running and that the CRUSH map is balanced. If it takes more than
two minutes, we fail the reconcile and try again. This is only done
the first time the stretch cluster is configured. In future reconciles
we first check if the stretch cluster is already enabled before
enabling it again.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
The mon timeouts and pod retries were removed back in v1.1,
this just removes some obsolete variables from the legacy
mon scheduling.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
The operator schedules all the mons at the same time, then
serially starts each of the mons. If the operator was restarted
or otherwise failed before all the mons were started, the next
reconcile will schedule new mons again and ignore the scheduling
that was already computed. Since the last mon running is always
expected to be in quorum, we can reset the maxMonId to be
the highest mon currently created.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
The arbiter may have unique requirements for placement settings. Thus, if the arbiter
placement is specified in the cluster CR, that placement will be applied to the arbiter
mon and the other mons will retain their placement from the other mon or all
sections of the placement.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
The node usage was used for the obsolete mon selection algorithm, which
was replaced long ago with placement via canary pods.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
We can now collect logs directly into a side-car container.
A new CRD spec has been added:
spec:
logCollector:
enabled: true
periodicity: 24h
Every 24h we will rotate log files for each Ceph daemon.
Signed-off-by: Sébastien Han <seb@redhat.com>
We can now prioritize orchestrations on certain event. Today only two
events will cancel on-going orchestrations (if any):
* request for cluster deletion
* request for cluster upgrade
If one of the two are caught by the watcher we will cancel the on-going
orchestration.
For that we implemented a simple approach based on check points, where
we will check for a cancellation request in certain part of the
orchestration. Mainly before each mons/mgr/osds orchestration loops.
This solution is not perfect, but we are waiting for the
controller-runtime to release its 0.7 version which will embed context
support. With that we will be able to cancel reconciles more precisely
and rapidly.
Operator log example:
```
2020-11-24 13:54:59.499719 I | op-mon: parsing mon endpoints: a=10.109.126.120:6789
2020-11-24 13:54:59.499719 I | op-mon: parsing mon endpoints: a=10.109.126.120:6789
2020-11-25 12:59:12.986264 I | ceph-cluster-controller: done reconciling ceph cluster in namespace "rook-ceph"
2020-11-25 13:07:33.776947 I | ceph-cluster-controller: CR has changed for "rook-ceph". diff= v1.ClusterSpec{
CephVersion: v1.CephVersionSpec{
Image: "ceph/ceph:v15.2.5",
- AllowUnsupported: true,
+ AllowUnsupported: false,
},
DriveGroups: nil,
Storage: {UseAllNodes: true, Selection: {UseAllDevices: &true}},
... // 20 identical fields
}
2020-11-25 13:07:33.777039 I | ceph-cluster-controller: reconciling ceph cluster in namespace "rook-ceph"
2020-11-25 13:07:33.785088 I | op-mon: parsing mon endpoints: a=10.107.242.49:6789,b=10.109.71.30:6789,c=10.98.93.224:6789
2020-11-25 13:07:33.788626 I | ceph-cluster-controller: detecting the ceph image version for image ceph/ceph:v15.2.5...
2020-11-25 13:07:35.280789 I | ceph-cluster-controller: detected ceph image version: "15.2.5-0 octopus"
2020-11-25 13:07:35.280806 I | ceph-cluster-controller: validating ceph version from provided image
2020-11-25 13:07:35.285888 I | op-mon: parsing mon endpoints: a=10.107.242.49:6789,b=10.109.71.30:6789,c=10.98.93.224:6789
2020-11-25 13:07:35.287828 I | cephclient: writing config file /var/lib/rook/rook-ceph/rook-ceph.config
2020-11-25 13:07:35.288082 I | cephclient: generated admin config in /var/lib/rook/rook-ceph
2020-11-25 13:07:35.621625 I | ceph-cluster-controller: cluster "rook-ceph": version "15.2.5-0 octopus" detected for image "ceph/ceph:v15.2.5"
2020-11-25 13:07:35.642688 I | op-mon: start running mons
2020-11-25 13:07:35.646323 I | op-mon: parsing mon endpoints: a=10.107.242.49:6789,b=10.109.71.30:6789,c=10.98.93.224:6789
2020-11-25 13:07:35.654070 I | op-mon: saved mon endpoints to config map map[csi-cluster-config-json:[{"clusterID":"rook-ceph","monitors":["10.107.242.49:6789","10.109.71.30:6789","10.98.93.224:6789"]}] data:a=10.107.242.49:6789,b=10.109.71.30:6789,c=10.98.93.224:6789 mapping:{"node":{"a":{"Name":"minikube","Hostname":"minikube","Address":"192.168.39.3"},"b":{"Name":"minikube","Hostname":"minikube","Address":"192.168.39.3"},"c":{"Name":"minikube","Hostname":"minikube","Address":"192.168.39.3"}}} maxMonId:2]
2020-11-25 13:07:35.868253 I | cephclient: writing config file /var/lib/rook/rook-ceph/rook-ceph.config
2020-11-25 13:07:35.868573 I | cephclient: generated admin config in /var/lib/rook/rook-ceph
2020-11-25 13:07:37.074353 I | op-mon: targeting the mon count 3
2020-11-25 13:07:38.153435 I | op-mon: checking for basic quorum with existing mons
2020-11-25 13:07:38.178029 I | op-mon: mon "a" endpoint is [v2:10.107.242.49:3300,v1:10.107.242.49:6789]
2020-11-25 13:07:38.670191 I | op-mon: mon "b" endpoint is [v2:10.109.71.30:3300,v1:10.109.71.30:6789]
2020-11-25 13:07:39.477820 I | op-mon: mon "c" endpoint is [v2:10.98.93.224:3300,v1:10.98.93.224:6789]
2020-11-25 13:07:39.874094 I | op-mon: saved mon endpoints to config map map[csi-cluster-config-json:[{"clusterID":"rook-ceph","monitors":["10.107.242.49:6789","10.109.71.30:6789","10.98.93.224:6789"]}] data:a=10.107.242.49:6789,b=10.109.71.30:6789,c=10.98.93.224:6789 mapping:{"node":{"a":{"Name":"minikube","Hostname":"minikube","Address":"192.168.39.3"},"b":{"Name":"minikube","Hostname":"minikube","Address":"192.168.39.3"},"c":{"Name":"minikube","Hostname":"minikube","Address":"192.168.39.3"}}} maxMonId:2]
2020-11-25 13:07:40.467999 I | cephclient: writing config file /var/lib/rook/rook-ceph/rook-ceph.config
2020-11-25 13:07:40.469733 I | cephclient: generated admin config in /var/lib/rook/rook-ceph
2020-11-25 13:07:41.071710 I | cephclient: writing config file /var/lib/rook/rook-ceph/rook-ceph.config
2020-11-25 13:07:41.078903 I | cephclient: generated admin config in /var/lib/rook/rook-ceph
2020-11-25 13:07:41.125233 I | op-mon: deployment for mon rook-ceph-mon-a already exists. updating if needed
2020-11-25 13:07:41.327778 I | op-k8sutil: updating deployment "rook-ceph-mon-a" after verifying it is safe to stop
2020-11-25 13:07:41.327895 I | op-mon: checking if we can stop the deployment rook-ceph-mon-a
2020-11-25 13:07:44.045644 I | op-k8sutil: finished waiting for updated deployment "rook-ceph-mon-a"
2020-11-25 13:07:44.045706 I | op-mon: checking if we can continue the deployment rook-ceph-mon-a
2020-11-25 13:07:44.045740 I | op-mon: waiting for mon quorum with [a b c]
2020-11-25 13:07:44.109159 I | op-mon: mons running: [a b c]
2020-11-25 13:07:44.474596 I | op-mon: Monitors in quorum: [a b c]
2020-11-25 13:07:44.478565 I | op-mon: deployment for mon rook-ceph-mon-b already exists. updating if needed
2020-11-25 13:07:44.493374 I | op-k8sutil: updating deployment "rook-ceph-mon-b" after verifying it is safe to stop
2020-11-25 13:07:44.493403 I | op-mon: checking if we can stop the deployment rook-ceph-mon-b
2020-11-25 13:07:47.135524 I | op-k8sutil: finished waiting for updated deployment "rook-ceph-mon-b"
2020-11-25 13:07:47.135542 I | op-mon: checking if we can continue the deployment rook-ceph-mon-b
2020-11-25 13:07:47.135551 I | op-mon: waiting for mon quorum with [a b c]
2020-11-25 13:07:47.148820 I | op-mon: mons running: [a b c]
2020-11-25 13:07:47.445946 I | op-mon: Monitors in quorum: [a b c]
2020-11-25 13:07:47.448991 I | op-mon: deployment for mon rook-ceph-mon-c already exists. updating if needed
2020-11-25 13:07:47.462041 I | op-k8sutil: updating deployment "rook-ceph-mon-c" after verifying it is safe to stop
2020-11-25 13:07:47.462060 I | op-mon: checking if we can stop the deployment rook-ceph-mon-c
2020-11-25 13:07:48.853118 I | ceph-cluster-controller: CR has changed for "rook-ceph". diff= v1.ClusterSpec{
CephVersion: v1.CephVersionSpec{
- Image: "ceph/ceph:v15.2.5",
+ Image: "ceph/ceph:v15.2.6",
AllowUnsupported: false,
},
DriveGroups: nil,
Storage: {UseAllNodes: true, Selection: {UseAllDevices: &true}},
... // 20 identical fields
}
2020-11-25 13:07:48.853140 I | ceph-cluster-controller: upgrade requested, cancelling any ongoing orchestration
2020-11-25 13:07:50.119584 I | op-k8sutil: finished waiting for updated deployment "rook-ceph-mon-c"
2020-11-25 13:07:50.119606 I | op-mon: checking if we can continue the deployment rook-ceph-mon-c
2020-11-25 13:07:50.119619 I | op-mon: waiting for mon quorum with [a b c]
2020-11-25 13:07:50.130860 I | op-mon: mons running: [a b c]
2020-11-25 13:07:50.431341 I | op-mon: Monitors in quorum: [a b c]
2020-11-25 13:07:50.431361 I | op-mon: mons created: 3
2020-11-25 13:07:50.734156 I | op-mon: waiting for mon quorum with [a b c]
2020-11-25 13:07:50.745763 I | op-mon: mons running: [a b c]
2020-11-25 13:07:51.045108 I | op-mon: Monitors in quorum: [a b c]
2020-11-25 13:07:51.054497 E | ceph-cluster-controller: failed to reconcile. failed to reconcile cluster "rook-ceph": failed to configure local ceph cluster: failed to create cluster: CANCELLING CURRENT ORCHESTATION
2020-11-25 13:07:52.055208 I | ceph-cluster-controller: reconciling ceph cluster in namespace "rook-ceph"
2020-11-25 13:07:52.070690 I | op-mon: parsing mon endpoints: a=10.107.242.49:6789,b=10.109.71.30:6789,c=10.98.93.224:6789
2020-11-25 13:07:52.088979 I | ceph-cluster-controller: detecting the ceph image version for image ceph/ceph:v15.2.6...
2020-11-25 13:07:53.904811 I | ceph-cluster-controller: detected ceph image version: "15.2.6-0 octopus"
2020-11-25 13:07:53.904862 I | ceph-cluster-controller: validating ceph version from provided image
```
Closes: https://github.com/rook/rook/issues/6587
Signed-off-by: Sébastien Han <seb@redhat.com>
In a disaster recovery scenario, the mon service may have been
accidentally deleted, while the expected mon endpoint is still
found in the mon endpoints configmap. In this case, we create
the mon service with the same endpoint as previously.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
Found by running the following command:
codespell -S .git,*.png,*.jpg -L \
aks,keyserver,atleast,dne,ser,ist,files\',ba,dum,iam,te -f -H
Signed-off-by: Mateusz Gozdek <mgozdekof@gmail.com>
Instead of directly defining our own constants, we should be using the k8s
constants for the well-known topology labels.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
Pools in stretch clusters will all use the same crush rule that
is generated by the operator when configuring the cluster for
stretch mode, with replica 4 and two replicas per failure domain.
Pools cannot create new crush rules, so we require that replica be
4 when the pools is created, EC is not allowed, and no new
crush rule will be created.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
In clusters where only two datacenters (or similar failure domains)
are available, a different mon and osd approach is needed to deal
with the network partitions or some other reason for one of the failure
domains going down. The Ceph stretched cluster makes the mons aware
of the failure domains by configuring one as the arbiter in a third
zone, while keeping two replicas of the data in each of the data
zoens.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
The mon canaries are now being added and removed at the same time
instead of starting or stopping them one-by-one. This will simplify
scheduling execution time to much more quickly detect where the mons
can be scheduled.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
The mon resources may be orphaned if there is a mon failover
that was cancelled, for example a mon pvc may remain if the
original mon comes up and the operator was restarted before
the mon failover completed. The cleaning up of mon resources
was only happening was each successful mon failover, now we
perform the check with every reconcile since the resources
were not being cleaned up predictably enough.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
golangci-lint linter gosec showing more errors
than gosec gh. This commit resolve
new errors. And, removing
nosec comments from autogenerated files.
Signed-off-by: subhamkrai <srai@redhat.com>
this commit handle golangci-lint linter errcheck.
`errcheck` - Errcheck is a program for checking for
unchecked errors in go programs. These unchecked errors
can be critical bugs in some cases
To see only staticcheck linter output
`golangci-lint run --disable-all -E errcheck`
Signed-off-by: subhamkrai <srai@redhat.com>
this commit handle golangci-lint linter staticcheck error.
`staticcheck` - Staticcheck is a go vet on steroids,
applying a ton of static analysis checks.
To see only `staticcheck` linter output
`golangci-lint run --disable-all -E staticcheck`
Signed-off-by: subhamkrai <srai@redhat.com>
this commit will enable one more linter ineffassign
in golangci-lint.
This linter throws an error when variable is assigned and never used.
`golangci-lint run --disable-all -E ineffassign` is used detects ineffassign
errors only.
Signed-off-by: subhamkrai <subhamkumarrai03@gmail.com>
The selector labels used for mons cannot change during upgrade
or else the upgrade will hang when the cluster is checking
for mon status immediately after the upgrade begins.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
If the mon failover fails, if the mon is on a pvc, the pvc could remain
orphaned. This is likely to happen if there is not a node available
for scheduling the mon. Therefore, the operator will ensure after a failover
that any orphaned mon pvc is cleaned up.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
When the mons are based on pvcs, the pvc_name label was not being set
as expected. Now the label will always be set on both the canary and
the daemon pods.
The mon pvcs were also never being deleted during canary cleanup. This
code is now removed since we must be very careful when the pvcs are
removed.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
This allows users to specify additional labels to be added to the components
created by the Rook Ceph operator.
Additionally the rbd mirrors are now getting their annotations and
labels set.
Signed-off-by: Alexander Trost <galexrt@googlemail.com>
The operator will now allow the resource limits to be applied below the
recommended minimums. In small clusters, even the recommended minimums
may not be necessary. A warning is still printed to the operator log,
but we allow the configuration to continue.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
When a mon is failed over, the new mon is started before the old mon
is removed from quorum and purged from the cluster. If the mon has
anti-affinity, in small clusters or specific toplogies where sufficient
nodes are not available, the new mon may be prevented from starting simply
because the dead mon pod still exists. To ensure a clean failover, the
dead mon pod must first be stopped before the new mon is started.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
For external cluster, it is not an issue to have a mon count set to 0
since Rook is not deploying/managing the monitors.
Signed-off-by: Sébastien Han <seb@redhat.com>
The service selector labels cannot be updated during the upgrade or else the
mon commands will not respond until the mon deployment is also updated.
Since there is a mon command to check for the ceph versions before the
deployment is updated, the operator would remain stuck during the upgrade.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
The labels cannot be updated on a deployment match selector.
In v1.4 a new label was added for ceph_daemon_type that is intended
to be on the pod labels, but cannot be applied to the match selectors.
Therefore, we suppress any new labels that are added to the daemons.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
Rename PodLabels function to CephDaemonAppLabels for more clarity about
what the function's purpose is.
Signed-off-by: Blaine Gardner <blaine.gardner@suse.com>
At startup there is a race condition between the goroutine checking
for mon health and the first reconcile of the mons. There is a lock
that ensures only one of them will check the mon health at a time.
The issue is that if the wrong one gets there first, it will not
have the desired mon count initialized from the cluster CR as expected
and the healther checker will use a desired mon count of 0.
The desired count of 0 leads the health checker to remove
a mon since there are extra mons. Only one mon is allowed
to be removed at a time, and adjusting to fewer than two mons is
disallowed once there are at least two mons.
After the health checker is completed, the main reconcile of the mons
will proceed with the correct desired mon count of 3 and attempt
to restore to the real desired quorum size.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
The mon was starting with a partial spec. There was a race condition
that was causing incorrect initialization in the health check goroutine
that could remove a mon upon operator restart. Now the mon will be
properly initialized with the full spec.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
The user creds should not be passed to the ceph helper methods
since the clusterInfo already has the needed context. Removing
a final reference that was missed in the refactor.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
The external creds have been loaded from their own key. For
backward compatibility we need to check that key if the
new key is not found.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
To provide more context for executing commands in a ceph cluster,
the full clusterInfo is now passed to the ceph execution commands.
All information about the cluster will now be available throughout
all the areas of the operator. The namespace, ceph credentials,
mon endpoints, and other info is a core part of that cluster info.
Arguments passed through the controllers are also simplified for
mons, mgr, osds, and other daemons where the parameters had
become too complex.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
The ceph config and client packages are conceptually the same.
To avoid circular dependencies in some cases, we simplify by combining
the packages into the client package.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
The operator should only connect to ceph with a single set of creds.
In a converged cluster this will be the admin creds and in an external
cluster it will be lower-privileged creds. Independent clusters were
implemented with a separate set of creds. To simplify the code these
are now merged to a single set.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
This commit allows us to configure status check for each daemon:
* "mon": health check on the ceph monitors (quorum)
* "osd": health check on the ceph osds
* "status": ceph health status check
Each check is controlled by the following settings:
* disabled: whether to disable the check (default: false)
* internal: interval to run the check
* timeout: only valid for mons, is the timeout for unresponsive mon
before failling over.
Example to disable the status health check:
```yaml
healthCheck:
daemonHealth:
status:
disabled: true
```
As part of that, pod's livenessprobe can now be configured via the
following settings:
* disabled: whether to enable or not
* probe: override the current probe in place by a new one
```yaml
healthCheck:
livenessProbe:
mon:
disabled: true
```
Closes: https://github.com/rook/rook/issues/5772
Signed-off-by: Sébastien Han <seb@redhat.com>
The stuck mons were only being force deleted in the mon health checker.
After an operator restart, the health check may not be executed
for 20 minutes until after the timeout during the main reconcile
which is holding onto a lock for mon reconcile. Now the stuck mon
will be deleted during the reconcile as well.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
If the mon is backed by a pvc, ensure the pvc is deleted when
the mon is failed over. The pvc is being left in a bound state
even though it will no longer be used in the cluster.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
Sébastien Han
Travis Nielsen
Arun Kumar Mohan
Mateusz Gozdek
Yusei Nishiyama
subhamkrai
Santosh Pillai
subhamkrai
Alexander Trost
Blaine Gardner