“Reclaimed” does not tell you by itself why an Oracle Cloud instance stopped, or whether its disk can be used again. Check the instance state and Oracle’s notification first; then choose a recovery path based on whether the instance, boot volume or a backup still exists.
If you recover a machine or create a replacement, that restores only part of the job. You still need to start the encoder, reconnect it to the correct YouTube live event, check its preview and confirm that viewers can see and hear the stream.
What a reclaim means—and what it does not
People often use “reclaimed” to describe any virtual machine that has disappeared or stopped running. Oracle’s console may show a more specific state, and a notification or event history may give more context. Use those details rather than treating “reclaimed” as a formal diagnosis or a recovery instruction.
A stopped instance and a terminated instance are not the same recovery problem. A stopped machine may be available to start again, subject to its current state and placement. A terminated instance has been deleted; recovery then depends on whether its boot volume or a suitable backup is still available. Neither the instance label nor the word “reclaimed” proves that data has been kept.
Preemptible capacity is one specific case. Oracle says that preemptible instances can be terminated when capacity is reclaimed and that they are not suited to long-running workloads. That is not evidence that every free-tier stop, termination or other interruption was preemption. Check the instance’s configuration and the notice Oracle sent before drawing that conclusion. Oracle’s preemptible-instance documentation describes that case.
For a channel that plays a bhajan loop or a local news bulletin around the clock, the practical distinction is whether you can start the same instance, launch a replacement using retained storage, or must rebuild from another copy of your files. Keep a separate copy of the original video and the encoder settings; a VM should not be the only place where either one exists.
Check the instance state and Oracle notifications
Open the Oracle Cloud console and select the right tenancy, region and compartment before looking for the instance. Confirm whether the state is stopped, terminated, disabled, or another unavailable state. A search in the wrong region can make a retained instance or volume appear to be missing, so verify the location shown in any Oracle notice as well.
Read the notification closely. Note the affected resource, the stated reason, any event time and whether Oracle directs you to take an action. If there is a relevant event-history entry, compare it with the notice. Preserve the exact wording for your own troubleshooting notes. Avoid relying on a shortened message or a screenshot that omits the instance name and region.
Then check the instance’s associated storage. Look for a boot volume in the same region and availability domain, and note its name and attachment state. If it is present, do not detach or delete it while you are still deciding on a recovery path. If it is absent, check whether a boot-volume backup exists and whether you are viewing the region in which that backup was stored.
A simple record helps avoid repeating work: instance state, region, availability domain, notification text, boot-volume state and any backup you can identify. Do not assume that a stopped state means preemption, or that a terminated state means the volume has also been deleted. Verify each resource separately. Oracle’s manual migration instructions cover creating an instance from a boot volume associated with a terminated instance, but they are not a promise that your particular volume remains available.
Identify preemptible-instance termination
If the notice or instance configuration identifies preemptible capacity, treat that as a meaningful explanation rather than a generic synonym for interruption. Oracle’s documentation says preemptible instances may be terminated at any time. A 24/7 encoder is a long-running workload, so a design that depends on such an instance continuing indefinitely carries an interruption risk by definition.
If you did not select a preemptible instance, do not infer that it was preemptible simply because it stopped or disappeared. Look for the actual configuration and Oracle’s message. Free-tier eligibility, service limits, an administrative action, a disabled resource and a preemptible termination are different possibilities; the available evidence should determine what you investigate next.
For the stream, write down what the interruption actually affected. A terminated compute instance may have taken the encoder process offline, but the YouTube live event and its settings are separate. Likewise, a live event that has ended is not made active just by bringing the VM back. Check the event in YouTube Studio before choosing whether to reconnect to it or prepare another broadcast.
If the instance was preemptible, first see whether the boot volume survived and whether a replacement can be placed. If you find no usable volume or backup, plan to rebuild from source video and documented settings rather than expecting Oracle to reinstate the deleted machine. Do not spend time changing unrelated stream settings until you know which compute and storage resources are actually recoverable.
Assess whether the boot volume remains available
A boot volume is the disk from which the instance’s operating system and installed applications start. If Oracle still lists the volume, it may let you use that disk as the basis for a replacement instance. That can preserve more of the installed system than starting from a clean image, but it does not guarantee that the replacement will have the same networking, attached data volumes or running applications.
If the original boot volume is unavailable, look for a boot-volume backup that was created and retained before the interruption. Oracle documents making a new boot volume from a backup for disaster recovery and migration. A backup is useful only if it exists, is accessible in the region you intend to use and contains the state you need. Check its age and any relevant retention details before relying on it. Oracle’s cross-region boot-volume backup guidance describes this route.
If neither the volume nor a usable backup exists, the evidence here does not support a promise that the terminated instance can be recovered. You may still be able to build a new encoder machine, but that is a rebuild, not a restoration of the old installation. This distinction matters if you stored the only copy of a playlist, script, licence file or audio asset on the VM.
| What you find | Likely next step | What still needs checking |
|---|---|---|
| Instance is stopped and remains available | Try starting it after reviewing its current state | Confirm the encoder starts and reconnects to the intended event |
| Instance is terminated, boot volume retained | Follow Oracle’s documented flow to create a replacement from that volume | Placement capacity, network configuration and attached resources |
| Original volume is missing, backup retained | Create a boot volume from the backup and use it for recovery | Backup availability, region and how much state it contains |
| No usable volume or backup | Rebuild from source files and recorded settings | YouTube event details, networking and application configuration |
Restore or create an instance from the volume
If the original instance is still present and stopped, use the console’s available start action and watch for an error rather than repeatedly clicking. If it is terminated, follow the current OCI instructions for creating a new instance from the retained boot volume. In either case, first check the selected region and availability domain and confirm the storage resource you intend to use.
A replacement made from a boot volume is not necessarily a complete copy of the old operating environment. Oracle notes that manual migration can require extra steps for instances with additional VNICs, secondary IP addresses, attached block volumes, TPM settings or load-balancer membership. If your stream depended on any of these, write down what must be recreated and check the relevant network and storage resources before testing the encoder.
When instance creation fails, identify whether the error is about a tenancy service limit or host capacity. Oracle distinguishes service limits from the physical availability of compute capacity. If the message says “out of host capacity”, changing a quota alone may not resolve it. You can review the applicable service limit and consider another eligible placement or shape, but Oracle does not promise that any particular region, availability domain or retry will have capacity. Its compute overview explains the distinction.
Do not delete the retained volume simply because a launch attempt failed. Keep your recovery notes and the volume or backup intact while you investigate the exact error. Once the new machine is available, verify its operating system, network access, clock, storage mounts and startup services before treating it as ready for a continuous broadcast.
Reinstall the encoder and restore stream settings
A replacement instance may preserve the encoder if it was created from the old boot volume, but check rather than assume. Start the application manually once, confirm it can read the media file and verify that its output device and audio source are correct. If you have to rebuild from a clean image, install the encoder again and restore the playlist, scene layout, scripts or other settings from your own saved copies.
Next, open the relevant live event in YouTube Studio’s Live Control Room. Check that it is the event you intend to broadcast to, then retrieve its current stream URL and stream key. The URL tells the encoder where to send its feed, while the key identifies the stream connection. Enter both in the encoder and keep the key private. YouTube’s live stream settings help explains where to manage them.
If you suspect the key was exposed, reset it in YouTube Studio and replace the saved value in the encoder. Do not paste a stream key into a public support post or an unredacted screenshot. If you reuse a saved configuration, check that it does not contain an old key or point to an unintended event.
Start the encoder and wait for YouTube’s preview before starting or resuming the broadcast. Check picture, sound, title and event status. A successful VM launch is not proof that YouTube is receiving a usable feed, and an encoder that reports “streaming” is not proof that viewers can see the intended output. YouTube’s encoder setup guidance covers the connection and preview process.
For a playlist-based station, also check that playback advances beyond its opening file and that sound continues through a transition. If you use OBS scenes or a playlist, a saved configuration and a tested recovery routine are valuable; for example, these OBS playlist steps after logging out of Windows address a different failure point but illustrate why startup behaviour should be tested. If an Indian channel recently moved to a Brand Account, confirm access separately using this guide to YouTube live access after a Brand Account move.
Reduce the impact of another interruption
A recovery is quicker when the parts of the stream that do not belong on the VM are stored somewhere you control. Keep original media, encoder profiles, playlist order, startup instructions and any required licences in a separate location. Keep a brief record of the YouTube event workflow, but store the stream key securely rather than in a plain document shared with helpers.
Decide what you will do if the instance stops again. That can mean checking Oracle’s notice and resource state, then starting the retained instance or launching from a backup if available. If you maintain a second encoder setup, test it before relying on it; failover is a plan to validate, not an automatic guarantee. YouTube recommends checking audio and video quality and testing backup encoder arrangements in its live-streaming tips.
For an always-on channel, a computer at home or an office may also be affected by power, broadband or a router restart. A playlist stream affected by an IP change is a separate failure mode, but the lesson is similar: distinguish a problem in the source machine or network from one in YouTube Studio before changing settings. Keep a written sequence that starts with checking the stream health and ends with confirming playback from a viewer’s device.
Do not rely on YouTube’s automatic archive as the sole recording of an uninterrupted day-long channel. YouTube’s encoder guidance says streams under 12 hours are automatically archived; that guidance does not mean a continuous 24/7 stream will necessarily appear as one complete recording. If aarti, a class or a news bulletin must be kept, plan a separate local or off-machine recording and check that it is actually being saved.
If repeated instance interruptions make operating a VM difficult, one alternative is to move the upload and stream operation away from your own always-on computer. StreamNeo removes that particular maintenance task by turning an uploaded video into a YouTube stream that continues with your computer switched off; it does not remove the need to check the channel, event and playback.
Before committing, compare the operating options on the pricing page. When the file and channel are ready, start free — 24-hour trial, no card.
FAQ
Does “reclaimed” mean Oracle terminated my instance?
Not by itself. “Reclaimed” is often used loosely, so check the actual OCI state and the notification or event history. Preemptible termination is one possible cause, not a conclusion to apply to every stopped or terminated instance.
Can I restore a terminated instance from its boot volume?
If the boot volume still exists, Oracle documents creating a new instance using the volume from a terminated instance. You may need to recreate networking and attached resources, and a retained volume does not guarantee compute capacity is available. If the volume is gone, check for a usable boot-volume backup; neither route guarantees recovery in every case.
Will restoring the VM make my YouTube stream resume?
No. You must start or restore the encoder, check the current event’s stream URL and key, and wait for YouTube’s preview. Confirm the output and playback before considering the channel live again.
Will YouTube keep a complete archive of a continuous 24/7 stream?
Do not assume so. YouTube’s guidance says streams under 12 hours are automatically archived, which does not promise one complete archive for a continuous 24/7 broadcast. Use a separate recording plan if you need a dependable copy.