The csi-yadro-tatlin-unified module: FAQ

How to check module health?

To do this, you need to check the status of the pods in the d8-csi-yadro-tatlin-unified namespace. All pods should be in the Running or Completed state and should be running on all nodes.

kubectl -n d8-csi-yadro-tatlin-unified get pod -owide -w

How much of a volume is reserved for the superuser?

Nothing, by default: a volume formatted with ext4 gets -m0, so the whole volume is available to the workload.

Where the classic 5% ext4 reserve is wanted — for example to keep a filesystem writable for a privileged process after a workload has filled it — annotate the YadroTatlinUnifiedStorageClass with the percentage to reserve:

d8 k annotate yadrotatlinunifiedstorageclass <name> storage.deckhouse.io/ext4-reserved-percent=5

The value is a whole number of percent between 0 and 50; an invalid one leaves the YadroTatlinUnifiedStorageClass with Ready=False and the reason in its status, instead of breaking volume creation later.

The reserve applies only to volumes created after the annotation was set — filesystems that already exist keep the reserve they were created with. Only the ext filesystems keep blocks for the superuser, so a class with fsType: xfs ignores it.

Changing the annotation makes the controller recreate the StorageClass, because parameters of an existing StorageClass are immutable in Kubernetes. Existing volumes and PVCs are not affected.

What happens if a node’s repositories cannot provide open-iscsi, multipath-tools or nvme-cli?

The module installs them itself, from images of its own.

On every node it serves, the NodeGroupConfiguration first asks the node’s package manager for open-iscsi and multipath-tools, and — where the module is configured for NVMe-TCP — for nvme-cli. That stays the preferred path: a host’s own client and daemon are what the rest of that host is built around, and nothing about them is the module’s business while they are there.

When an install fails — a closed environment with no route to the repositories, or a distribution that does not ship the packages — the module falls back to a package image of its own. bashible pulls it from the module’s registry, unpacks it on the node and runs its install script.

There are two such images, and they are independent on purpose: a distribution that cannot provide open-iscsi may still provide nvme-cli, or the other way round.

iscsi-tools lands under /var/lib/deckhouse/sds/csi-yadro-tatlin-unified and brings up two units:

systemctl is-active d8-csi-yadro-tatlin-unified-iscsid.service d8-csi-yadro-tatlin-unified-multipathd.service

nvme-tools lands under /var/lib/deckhouse/sds/csi-yadro-tatlin-unified-nvme and carries nvme alone — no daemon, and so no version to match. The host needs it for its own configuration rather than for the driver’s calls: the NodeGroupConfiguration runs nvme gen-hostnqn to give the node a stable NVMe Qualified Name, which the driver then registers in the array’s access group. Without it the driver reports initiators is empty and no NVMe-TCP volume can be published to that node.

Two things the iscsi-tools install script arranges that are worth knowing about, because both are invisible until something does not mount. The driver reaches the host’s iscsiadm through the container’s /bin/iscsiadm, which chroots into /host and looks the client up there against a PATH of its own — the script leaves a wrapper at /usr/local/sbin/iscsiadm for it. And multipathd from the package reads its configuration under the prefix it was built with, so the script links multipath.conf and multipath from that prefix onto the node’s own, which is how the module’s Tatlin device block reaches the daemon.

The two paths never mix. A host that already has iscsiadm of its own keeps everything of its own, and the install script does nothing there: the client and the daemon must always come from the same source, because iscsiadm of one version does not speak to iscsid of another. The same holds for nvme.

To tell which path a node took, look at what is running on it:

# the distribution's stack
systemctl is-active iscsid multipathd
# the module's own
ls /var/lib/deckhouse/sds/csi-yadro-tatlin-unified/bin
ls /var/lib/deckhouse/sds/csi-yadro-tatlin-unified-nvme/bin
command -v iscsiadm nvme

Two things the fallback deliberately does not change. It never merges anything into the host’s /lib64 or /usr/lib — the payloads are self-contained. And it does not touch /etc/iscsi/initiatorname.iscsi or /etc/nvme/hostnqn when the node already has them: those names are the node’s identity on the array, registered there, and a node that comes back under a different one is a node the array has never heard of.