Dell OS10 Factory Reset: Two Methods and Post-Reset Setup

Dell PowerSwitch switches running SmartFabric OS10 do not offer a single "factory reset" command the way OS9 did, which forces administrators to learn two manual procedures: a quick software reset that clears the startup configuration, and a deeper wipe that removes the startup XML plus the Redis database that holds the running state. This guide walks through both methods, explains when each is appropriate, and covers what the switch looks like after the reset - default credentials, default STP mode, and the management interface state - so reconfigure goes smoothly.

What Happens at Factory Defaults

After a successful reset the switch boots with: default hostname OS10, username and password admin/admin, SSH enabled and Telnet disabled, spanning tree in RPVST+ mode, only the default VLAN 1 present, and the out-of-band management interface configured for DHCP. All VLT domains, port-channels, VLANs and user configurations are gone.

It is worth being precise about that last point, because "factory default" is not the same as "as shipped". A unit that left the Dell factory has an ONIE installer environment and no OS10 image installed at all. Resetting OS10 to factory defaults keeps the installed image and the license state, and only clears configuration. If you need the switch to behave like a brand-new unit - ONIE discovery prompt, no license, no image - the reset procedure here is not sufficient and you must reinstall through ONIE instead.

Before You Reset: What Gets Destroyed

Neither method is reversible, so capture the running state first. The table below is the minimum you should collect, and the commands produce plain text you can paste into a ticket:

OS10# show running-configuration | no-more
OS10# show running-configuration vlt
OS10# show vlan
OS10# show running-configuration spanning-tree
OS10# show version
OS10# show boot
OS10# show license status
OS10# show interface status | no-more

The VLT domain settings, VLAN-to-port mappings and STP configuration are the pieces you will need when reconfiguring the switch for its new role. Also record the management IP, default gateway and any AAA or TACACS+ server definitions, because those are precisely the things that make the switch unreachable after a reset if you forget them.

Method 1: delete startup-configuration and reload

This is the quickest official method and works from an SSH, Telnet or console session - just be aware that the reload immediately ends a remote session, so console access is safer:

OS10# delete startup-configuration
Proceed to delete startup-configuration [confirm yes/no(default)]: yes

OS10# reload
System configuration has been modified. Save? [yes/no]: no
Proceed to reboot the system? [confirm yes/no]: yes

Answer no to the save prompt so the running configuration is not written back before reboot. This detail trips people up constantly: if you answer yes, the switch writes the running configuration back into the startup file you just deleted, and the reload achieves nothing. You will boot back into a fully configured switch and wonder why the reset did not work.

Method 1 is the right choice when the switch is being handed to another team and you mainly want the configuration gone, or when you are troubleshooting a config corruption that prevents normal boot behaviour.

Method 2: Full Linux-Kernel Wipe

When you need to be certain that every trace of the configuration is gone - for example before returning a switch to stock or re-purposing it - delete the startup file and the Redis database from the underlying Linux environment. This method must be done from the console, because once the files are deleted you cannot log in remotely:

OS10# system bash
admin@OS10:/home/admin$ sudo -i
[sudo] password for admin: <admin password>
root@OS10:~# ls -ls /config/etc/opt/dell/os10/db_init/
root@OS10:~# ls -ls /config/var/lib/redis/

root@OS10:~# rm -rf /config/etc/opt/dell/os10/db_init/startup.xml
root@OS10:~# rm -vf /config/var/lib/redis/dump.rdb

root@OS10:~# sync
root@OS10:~# reboot -f

Double-check that both files are gone before rebooting. The switch comes back with a completely clean configuration.

The reason this method is more thorough than Method 1 is the Redis key-value store. OS10 keeps much of its operational state - interface counters, LLDP neighbours, some table state - in Redis rather than in the flat XML configuration file. Deleting only startup.xml clears the intended configuration but leaves whatever Redis has persisted. Deleting dump.rdb as well removes that residue, which is why this is the procedure Dell recommends when a switch is leaving the organisation.

Note the two paths involved: /config/etc/opt/dell/os10/db_init/startup.xml is the human-readable startup configuration, and /config/var/lib/redis/dump.rdb is the Redis snapshot. The sync before reboot -f matters - it flushes filesystem buffers so the deletion is committed before the forced reboot, rather than being rolled back by an unclean shutdown.

If you want a middle ground, Method 1 paired with a manual delete startup-configuration and then explicitly checking for orphaned Redis state is what most field engineers actually do. When in doubt before an RMA, use Method 2.

Which Method Should You Use?

• Method 1 - routine re-provisioning, handing a switch to another team, cleaning up a lab unit. Fast, works over SSH, one reboot.
• Method 2 - RMA, decommissioning, returning a lease unit, or any time you cannot afford residual state. Console only, guaranteed clean.
• Neither - if the goal is to recover a switch that will not boot, or to change the installed OS10 version, use the ONIE reinstall path covered in our Dell OS10 firmware upgrade guide.

Verifying the Reset

OS10# show vlan
OS10# show running-configuration vlt
OS10# show interface port-channel summary
OS10# show spanning-tree active
OS10# show users

Only the default VLAN 1 should exist, the VLT and port-channel outputs should be empty, and spanning tree should show rapid-pvst as the active mode. show users should list only the admin account you are logged in with. If any local user you created earlier is still present, the reset did not fully take and you should repeat Method 2.

One more check that is easy to forget: confirm the interface names have returned to their defaults and that any breakout or split-port configuration is gone. A leftover breakout setting can make a 100G port appear as four 25G ports and confuse the next engineer considerably.

Post-Reset Management Configuration

The out-of-band management interface (mgmt1/1) defaults to DHCP. To configure it statically and restore management reachability:

OS10# configure terminal
OS10(config)# interface mgmt1/1
OS10(config-if-mgmt1/1)# no ip address dhcp
OS10(config-if-mgmt1/1)# ip address 192.168.1.2/24
OS10(config-if-mgmt1/1)# exit
OS10(config)# management route 0.0.0.0/0 192.168.1.1
OS10(config)# end
OS10# write memory

The management route command is the piece people miss. OS10 separates the management VRF from the default VRF, so an ip route 0.0.0.0/0 statement does not give the management interface a path off-box. Without the management route, you can ping the switch on its management IP from the same subnet but you cannot reach it from anywhere else, and the switch itself cannot reach a TACACS+ or NTP server.

Rebuilding the Switch in the Right Order

Once management is reachable, rebuild deliberately rather than pasting an old configuration wholesale - half the value of a reset is a clean slate:

OS10(config)# hostname DC1-LEAF-03
OS10(config)# username netops password <password> role sysadmin
OS10(config)# interface mgmt1/1
OS10(config-if-mgmt1/1)# no ip address dhcp
OS10(config-if-mgmt1/1)# ip address 10.10.0.13/24
OS10(config)# management route 0.0.0.0/0 10.10.0.1
OS10(config)# vlan 10,20,30
OS10(config)# interface ethernet1/1/1
OS10(config-if)# no switchport
OS10(config-if)# ip address 10.0.0.1/31
OS10(config)# spanning-tree mode rapid-pvst
OS10(config)# end
OS10# write memory

Bring up management first, then VLANs and access ports, then routed uplinks, then VLT last. VLT is deliberately sequenced at the end because it depends on the peer-link and the management reachability of both peers; configuring it before the underlay is stable creates a situation where one peer is up, the other is not, and the domain will not come into service cleanly.

If the switch is destined for an automated fabric, consider skipping manual rebuild entirely and using the OS10 zero-touch provisioning workflow described in our OS10 ZTD deployment guide - a factory-default switch is exactly the starting point ZTD expects.

Common Problems After a Reset

• Cannot reach the switch at all. The default mgmt1/1 setting is DHCP; if there is no DHCP server on the management VLAN the interface stays unconfigured. Connect to the console and set a static address.
• Login fails. Credentials are admin / admin. If that fails you are probably still talking to an old session or the reset did not complete.
• Console shows nothing. OS10 console defaults are 115200 baud, 8 data bits, no parity, 1 stop bit, no flow control. Legacy terminal profiles set to 9600 will show garbage.
• SSH refused but Telnet works. SSH is enabled and Telnet disabled by default on OS10; if neither works, check that your management route and an ACL are not blocking port 22.
• Config reappears after reboot. You answered "yes" at the save prompt in Method 1, or you skipped write memory after rebuilding and lost the new configuration on a power cycle.
• STP flooding or loops. RPVST+ is the default mode after a reset, which will not match a neighbour running MSTP. See our OS10 spanning-tree guide to align the region.

Frequently Asked Questions

Does a factory reset remove the OS10 image? No. The image stays in the alternate boot slot and a normal boot proceeds as before. Only the configuration and operational state are cleared.

Does it clear the license? No, license state is persisted separately and survives both reset methods. Check it with show license status if you have purchased features such as VLT or BGP.

Can I reset a stack or VLT pair in one go? Not with a single command. Each member is reset individually, and because the peer-link breaks when the first member reboots, take both members down together or accept that VLT will be out of service during the process. Verify the peer-link state afterwards with our VLT troubleshooting guide.

Is there a way to reset without a reboot? No. delete startup-configuration only removes the file on disk; the running configuration stays active until the switch reloads.

How long does the reset take? Typically three to five minutes for the reload plus a few more for the OS10 services to come up. Do not interrupt power.

Related reading: Dell OS10 factory reset by deleting config files from the console, Dell OS10 factory reset: delete startup-configuration and Dell OS10 default gateway and management route configuration.

Original article: Dell KB 000212446: How to Reset Dell OS10 Switch back to factory default | Dell InfoHub: Reset OS10 switches to factory defaults