Recovery Manager: Difference between revisions

From OSNEXUS Online Documentation Site
Jump to navigation Jump to search
Newer edit →
VisualWikitext
Create: dedicated Recovery Manager page, moved out of the Storage System page and expanded with the backup rotation, retention settings and cautions
 
m Trim and add a scope note: the Recovery Manager currently recovers osn.db only; a fuller configuration download/recovery workflow is planned, so this page deliberately stays brief
Line 2: Line 2:


The '''Recovery Manager''' restores the QuantaStor internal configuration
The '''Recovery Manager''' restores the QuantaStor internal configuration
database from a backup. Use it when a system has been re-installed and you need
database (<code>osn.db</code>) from a backup. Use it when a system has been
the previous configuration back -- Storage Volumes, Network Shares, Users,
re-installed and you need the previous configuration back -- Storage Volumes,
Hosts, Schedules, Roles, and the other logical objects.
Network Shares, Users, Hosts, Schedules and the other logical objects.


'''Navigation:''' Storage Management --> Storage Systems --> ''select a Storage System'' --> Recovery Manager ''(toolbar)''
'''Navigation:''' Storage Management --> Storage Systems --> ''select a Storage System'' --> Recovery Manager ''(toolbar)''
Line 11: Line 11:


[[File:ss_recovery_manager.png|thumb|center|800px|Recovery Manager, listing the configuration database backups discovered on the started Storage Pools.]]
[[File:ss_recovery_manager.png|thumb|center|800px|Recovery Manager, listing the configuration database backups discovered on the started Storage Pools.]]
'''Scope note:''' the Recovery Manager currently recovers the
<code>osn.db</code> configuration database only. A full configuration download
and recovery workflow covers considerably more than this, and is planned for a
future release. Treat this page as covering the database recovery step, not
configuration recovery as a whole.


'''Your data does not depend on this database.''' Storage Pools, and the volumes
'''Your data does not depend on this database.''' Storage Pools, and the volumes
and shares within them, carry their own metadata and are recovered by importing
and shares within them, carry their own metadata and are recovered by importing
the pool -- not from this backup. The configuration database holds the logical
the pool. The configuration database holds the logical objects layered on top,
objects layered on top. That is why a pool imported into a freshly installed
which is why a pool imported into a freshly installed system comes back with its
system comes back with its volumes intact but without the hosts, schedules and
volumes intact but without the hosts, schedules and users that referenced them.
users that referenced them.


'''In a Storage Grid the Recovery Manager matters much less''', because the
'''In a Storage Grid this matters much less''', because the configuration
configuration database is replicated across all grid nodes. A node rejoining a
database is replicated across all grid nodes; a node rejoining a grid picks the
grid picks the configuration up from its peers.
configuration up from its peers.


== Using the Recovery Manager ==
== Using the Recovery Manager ==


# Select the '''Storage System''' to recover.
# Select the '''Storage System''' to recover.
# Choose a backup from the list. Each row shows the backup name, its file size, and when it was created.
# Choose a backup from the list. Each row shows its name, file size and creation time.
# Optionally tick '''Include network configuration recovery''' -- see the warning below before you do.
# Optionally tick '''Include network configuration recovery''' -- see the cautions below.
# Press '''OK'''.
# Press '''OK'''.


The list is built from the backups discovered on the '''started''' Storage
The list is built from backups found on the '''started''' Storage Pools, so if
Pools. A pool that is not imported and started contributes nothing to the list,
the backup you expect is missing, import and start the pool holding it first.
so if the backup you expect is missing, import the pool that holds it first.
 
== How the backups are made ==


QuantaStor writes a copy of the configuration database to every Storage Pool on
Backups are named by rotation slot -- <code>osn.db.backup.hourly.N</code> and
a rotation, so the backups travel with the storage rather than living only on
<code>osn.db.backup.daily.N</code>. The count of each is set by
the boot drive. Backups are named by their rotation slot:
<code>max_db_backups</code> in
<code>/opt/osnexus/quantastor/conf/quantastor.conf</code> (default '''5'''; 0
disables pool backups). A backup failure raises a '''Configuration DB Backup
Failure''' alert -- see [[Call-home / Alerting|Alert Manager]].


* <code>osn.db.backup.hourly.1</code> through <code>osn.db.backup.hourly.N</code>
== Cautions ==
* <code>osn.db.backup.daily.1</code> through <code>osn.db.backup.daily.N</code>


<code>N</code> is controlled by <code>max_db_backups</code> in
'''Do not restore a backup taken from a different system.''' The database
<code>/opt/osnexus/quantastor/conf/quantastor.conf</code>, which defaults to
carries the system's unique ID, so this can leave two systems claiming the same
'''5''' of each. Values above 100 are capped at 100, and setting it to 0
UUID, which causes grid problems.
disables the pool backups entirely. The local copy on the boot drive is written
to:


<pre>
'''Network configuration recovery can cut off your own session.''' If the
/var/opt/osnexus/quantastor/osn.db.backup
restored network settings differ from the current ones, your connection drops
</pre>
when the address changes. Where ports may become unreachable, make the change
from the console rather than the web interface.


...on a cadence set by <code>hours_per_backup</code> in the same file, which
'''The restore replaces the running configuration.''' If you are unsure which
defaults to '''12''' hours.
backup is correct, contact OSNEXUS support at support@osnexus.com before
 
restoring rather than working through them by trial.
A failure to write a backup raises a '''Configuration DB Backup Failure''' alert,
so a system that has quietly stopped producing backups is visible in the Alert
Manager rather than only discovered when a recovery is attempted. See
[[Call-home / Alerting|Alert Manager]].
 
Because the backup is an ordinary file, it can also be replicated off the
appliance with <code>rsync</code> or <code>sftp</code> from a cron job if you
want a copy outside the storage.
 
== Important cautions ==
 
'''Recovering a database from a different system creates a duplicate identity.'''
The database carries the system's unique ID. Restoring a backup taken from
another server can leave two systems claiming the same UUID, which causes
problems in a grid. Only restore a backup that came from the system you are
restoring onto, unless OSNEXUS support has advised otherwise.
 
'''Network configuration recovery can cut off your own session.''' Ticking
'''Include network configuration recovery''' restores the saved network
settings, and if those differ from the current ones your connection drops when
the address changes. Where all network ports may become unreachable, make the
change from the console instead of the web interface, so you retain access.
 
For context, a newly installed QuantaStor system defaults to DHCP on an
interface. If a DHCP server is present it takes an automatic address, giving you
initial access from which to assign static addresses.
 
'''Restoring is not reversible in place.''' The restored database replaces the
running configuration. If you are uncertain which backup is the right one,
contact support before restoring rather than working through them by trial.


== Restoring by hand ==
== Restoring by hand ==
Line 97: Line 70:
systemctl start quantastor
systemctl start quantastor
</pre>
</pre>
== Getting help ==
Database recovery is one of the operations where it is worth asking first. If
you are unsure which backup to use, or a recovery has not produced what you
expected, contact OSNEXUS support at support@osnexus.com.


== Related pages ==
== Related pages ==


* [[Storage System]] -- system-level configuration and the Modify dialog
* [[Storage System]] -- system-level configuration and the Modify dialog
* [[Upgrade Manager]] -- includes a full re-install / boot drive swap procedure that uses the Recovery Manager
* [[Upgrade Manager]] -- includes a re-install / boot drive swap procedure that uses the Recovery Manager
* [[Call-home / Alerting|Alert Manager]] -- the Configuration DB Backup Failure alert
* [[Call-home / Alerting|Alert Manager]] -- the Configuration DB Backup Failure alert

Revision as of 10:37, 2 September 2026


The Recovery Manager restores the QuantaStor internal configuration database (osn.db) from a backup. Use it when a system has been re-installed and you need the previous configuration back -- Storage Volumes, Network Shares, Users, Hosts, Schedules and the other logical objects.

Navigation: Storage Management --> Storage Systems --> select a Storage System --> Recovery Manager (toolbar)

It is also on the right-click context menu for a Storage System.

Recovery Manager, listing the configuration database backups discovered on the started Storage Pools.

Scope note: the Recovery Manager currently recovers the osn.db configuration database only. A full configuration download and recovery workflow covers considerably more than this, and is planned for a future release. Treat this page as covering the database recovery step, not configuration recovery as a whole.

Your data does not depend on this database. Storage Pools, and the volumes and shares within them, carry their own metadata and are recovered by importing the pool. The configuration database holds the logical objects layered on top, which is why a pool imported into a freshly installed system comes back with its volumes intact but without the hosts, schedules and users that referenced them.

In a Storage Grid this matters much less, because the configuration database is replicated across all grid nodes; a node rejoining a grid picks the configuration up from its peers.

Using the Recovery Manager

  1. Select the Storage System to recover.
  2. Choose a backup from the list. Each row shows its name, file size and creation time.
  3. Optionally tick Include network configuration recovery -- see the cautions below.
  4. Press OK.

The list is built from backups found on the started Storage Pools, so if the backup you expect is missing, import and start the pool holding it first.

Backups are named by rotation slot -- osn.db.backup.hourly.N and osn.db.backup.daily.N. The count of each is set by max_db_backups in /opt/osnexus/quantastor/conf/quantastor.conf (default 5; 0 disables pool backups). A backup failure raises a Configuration DB Backup Failure alert -- see Alert Manager.

Cautions

Do not restore a backup taken from a different system. The database carries the system's unique ID, so this can leave two systems claiming the same UUID, which causes grid problems.

Network configuration recovery can cut off your own session. If the restored network settings differ from the current ones, your connection drops when the address changes. Where ports may become unreachable, make the change from the console rather than the web interface.

The restore replaces the running configuration. If you are unsure which backup is correct, contact OSNEXUS support at support@osnexus.com before restoring rather than working through them by trial.

Restoring by hand

The database is a SQLite file and can be replaced directly if the dialog is not available -- for example on a system whose web interface will not start: 

systemctl stop quantastor
cp <backup-file> /var/opt/osnexus/quantastor/osn.db
systemctl start quantastor

Related pages

  • Storage System -- system-level configuration and the Modify dialog
  • Upgrade Manager -- includes a re-install / boot drive swap procedure that uses the Recovery Manager
  • Alert Manager -- the Configuration DB Backup Failure alert
Retrieved from "https://wiki.osnexus.com/index.php?title=Recovery_Manager&oldid=27287"