How to Recover Swarm Cluster Services (SCS) Without a Backup File

Prerequisites

• The network interface was originally used for the private network, which the running cluster uses. The SCS instance will become the network gateway for the running storage cluster. Hence, there is a need for consistency between network settings for the recreated SCS instance and the running cluster.

  • The static IP for that network interface

  • The netmask for that network interface

• The Swarm storage administrator user name for the running storage cluster

• The Swarm storage administrator password for the running storage cluster

• The cluster name for the running storage cluster

• The DSN server(s) used by the running cluster

• The network domain name used for the running cluster

• The IP address pool reserved ranges are used when booting the running cluster.

Steps for Recovery

Step 1

Install SCS by following the instructions in https://perifery.atlassian.net/wiki/spaces/public/pages/3042345098/Online+SCS+Installation.

Step 2

Run the SCS initialization wizard by following the instructions in Run the Swarm Cluster Services (SCS) Initialization Wizard.

When prompted, use network settings (interface, IP address, and netmask) that are consistent with the running storage cluster.

Step 3

Install the Swarm storage component bundle.

• Add the Swarm storage component to the SCS repo.

# scsctl repo component add -f [storage software bundle file name]

• When asked whether to use encryption-at-rest, use the same mode that is consistent with the running cluster.

Missing setting: storage/disk.encryptNewVolumes

The following setting(s) require a default value:

disk.encryptNewVolumes [type: bool] (Whether to encrypt new Swarm volumes.  Enabling encryptNewVolumes means that any newly-formatted Swarm volume will be encrypted)
  True
  False

• When asked whether to enable multicast, be consistent with the multicast mode currently used in the running cluster.

The following setting(s) require a default value:

Missing setting: storage/cip.multicastEnabled

cip.multicastEnabled [type: bool] (Whether multicast should be used for communication within Swarm.)
  True
  False

• When asked for the name of the storage cluster, it is important to enter the same string as the name of the running cluster.

added: storage - 15.3.0 (15.3.0) [debug]

 Please enter a name for the group (FQDN format encouraged) - Required to be able to boot instances of "storage":

Step 4

Add additional settings.

# scsctl diagnostics config scan_missing

• In every case, when prompted, enter a value that is consistent with the settings in the running cluster.

• DNS servers:

----------------------------------------------
Name: network.dnsServers
Type: array[str]
Description: Required. DNS servers to be used.
----------------------------------------------
Missing setting: network_boot/network.dnsServers

 network.dnsServers [type: array[str]] (Required. DNS servers to be used.):

• Storage cluster administrator password: Enter the password being used in the running storage cluster.

Missing setting: platform/admin.password

admin.password [type: str ***SECURE***] (Administrative user password):

• DNS domain:

Missing setting: platform/network.dnsDomain

 network.dnsDomain [type: str] (Required. The DNS domain name that will be used.):

Step 5

Login with valid administrator credentials.

# scsctl auth login --user [administrator user name]:[administrator password]

Step 6

Storage cluster administrator user name:

• Proceed to step 8 If the Swarm storage administrator user name in the running cluster is the default value of “admin”.

• Change the administrator user name in SCS to the value in the running cluster. if the Swarm storage administrator user name in the running cluster is not the default value of “admin”.

# scsctl platform config set -d admin.userName=[admin user name in the running cluster] --allow-notification-failures

• You must login again after changing the administrator user name.

Step 7

Login with valid administrator credentials.

# scsctl auth login --user [administrator user name]:[administrator password]

Step 8

Initialize DHCP using the same reserved range parameters used originally for the storage cluster.

Subnet Layout:

   |                 |                                 |                 |
   |     reserve     |                                 |     reserve     |
   | <--  lower  --> | <---  storage pool range  ----> | <--  upper  --> |
   |      range      |                                 |      range      |
   |                 |                                 |                 |
 subnet              | <------ DHCP and Static ------> |             broadcast
 address                                                              address

Note that at least one of the reserved ranges must be set (lower or upper).

Step 9

Reboot the storage cluster in order for SCS to record the IP addresses of the storage cluster nodes (chassis).