Downloader

The Downloader is the node-local library for media and images used by Sylve. Add an installation ISO, a cloud-init image, a jail base archive, a disk image, or another file once, then select it from the workflows that support it.

Transfers and post-processing run in the background. The table refreshes while an item is unfinished, so it can show download progress, processing, completion, or the error returned by the source or converter.

Choose a source

Select New to add a source. The same field accepts a magnet link, an HTTP or HTTPS URL, or an absolute path that already exists on the Sylve node.

Source	Use it for	What Sylve does
HTTP or HTTPS URL	Downloading a file from a web server, such as an ISO or a VM image.	Retrieves the file into the Downloader library. You may provide a destination filename and enable post-processing.
Magnet link	Fetching a multi-file torrent, such as a distribution image.	Starts a BitTorrent transfer. Sylve adds its tracker list to the magnet link. Torrent entries can be viewed, but their source, name, and processing settings cannot be edited in the Downloader.
Absolute path	Bringing an existing regular file on the node into the Downloader library.	Copies the source file into the library. The source must be readable and must be a file, not a directory.
Upload	Sending files from the computer running your browser to the node.	Stages the upload first, then adds it to the same background Downloader workflow as any other source.

For an HTTP(S) source or a server-local path, Optional File Name controls the filename in the Downloader library. If it is blank, Sylve derives a name from the URL or path. Use a plain filename only, without directories.

Understand Download Type

Download Type classifies a completed item for other Sylve forms. It does not change the protocol used to retrieve the file.

Download Type	Choose it when	Where it matters
Uncategorized (ISOs, IMGs, etc.)	The file is general-purpose installation media, a disk image, or something with no special Sylve role.	Suitable for ordinary ISO and image selection workflows.
Base / RootFS	The file contains a jail base system or root filesystem.	Makes the item available as jail base or root filesystem media. Selecting it automatically enables extraction.
Cloud-Init	The file is a cloud-init capable VM image.	Makes the image eligible for a VM’s Cloud-Init media selection. A regular Base / RootFS or Uncategorized download cannot be used for that VM option.

Tip

For a VM image that needs Cloud-Init, select Cloud-Init when you add or upload it. The image must also be a cloud-init enabled operating-system image from its publisher, not merely an arbitrary ISO or disk image.

Transfer and processing options

Option	Available for	Effect
Ignore TLS Errors	HTTP(S) sources	Continues even when the remote server’s TLS certificate cannot be verified. Leave this off unless you understand and trust the source despite its certificate problem.
Extract Automatically	HTTP(S), node-local paths, and uploads	Extracts a tar archive or decompresses a supported compressed file after the file arrives. Base / RootFS enables this automatically.
Auto-convert to RAW	HTTP(S), node-local paths, and uploads	Converts a disk image to RAW using qemu-img. This is useful when an image is supplied as a format such as qcow2 or VMDK and the next workflow expects a RAW disk image.

Automatic extraction and RAW conversion can be enabled together for a compressed disk image. Sylve decompresses it first and converts the resulting image. Do not enable both for a tar archive such as .tar, .tar.gz, .txz, or .tar.zst: a tar archive is extracted as files and cannot be converted to a RAW disk image.

Torrent downloads do not support automatic extraction or RAW conversion. Download the torrent as-is, or obtain the image by HTTP(S), local path, or upload when post-processing is required.

Add a web, magnet, or local-path source

1. Open Utilities → Downloader and select New.
2. Paste a magnet link, an HTTP(S) URL, or the absolute path to a readable file on this node.
3. For an HTTP(S) or path source, optionally set a filename, then choose its Download Type.
4. Enable extraction or RAW conversion only when the file and intended use call for it.
5. Select Download.

The new row is accepted as Pending. It moves through transfer and post-processing without keeping the dialog open.

Table state	Meaning
Pending	Sylve accepted the item and has queued its work.
Processing	The transfer or its extraction/conversion work is running.
Done	The item is ready for selection by supported Sylve workflows.
Failed	The transfer or processing step stopped. Select the row’s progress cell to copy the reported error, correct the source or option, and add it again if needed.

Upload files from your computer

Select Upload in the upper-right corner. You can add multiple browser-local files to one staging queue. An upload is not yet a Downloader item until you complete it.

1. Drop files into the upload area or browse for them.
2. Set Batch defaults for the group, then select Apply to unfinished if you want to copy those choices to every staged file.
3. Review or change the Upload Type, extraction, and RAW conversion options for each individual file.
4. Select Complete for one item or Complete All for every valid staged item.

The browser transfer and completion are separate steps. This lets you classify and process each uploaded file before it enters the library. You can cancel unfinished items. Closing the dialog also clears unfinished staged uploads; files already completed remain in the Downloader.

Note

Uploaded files are staged temporarily before completion. Complete an upload while you are working, rather than treating staging as permanent storage.

Upload limits

Browser uploads use the node’s uploads section in config.json. These limits are enforced by the server, so changing browser behavior cannot bypass them.

config.json setting	Default	Effect
uploads.maxFileBytes	68719476736 bytes, or 64 GiB	Maximum size of each uploaded file. The multipart request has a small additional allowance for upload metadata.
uploads.maxConcurrentTransfers	2	Maximum number of upload transfers that the node accepts at once. This shared limit covers both Downloader uploads and File Explorer uploads. Further transfers wait for capacity and may need to be retried.

Set both values to positive integers. When a value is omitted or set to zero, Sylve uses its default. Consider available disk space in the node’s data path as well as the final image size, since a file is staged before it is completed into the Downloader library.

Reuse and manage completed items

Select a completed item in the table to reveal the available actions.

Action	Availability	Result
Download	A completed HTTP, local-path, uploaded, or individual torrent file row	Downloads that file to the computer running your browser.
Copy URL	The same completed selections as Download	Copies a short-lived, signed browser download URL. Treat it as temporary access.
Edit	Completed HTTP, local-path, and uploaded items	Change the display name, Download Type, and unfinished processing options. Enabling extraction or conversion later queues processing again. The source URL cannot be changed.
View	Torrent entries	Shows the torrent’s name and magnet link.
Delete	One or more top-level items	Removes the Downloader record and its managed data after safety checks.

For torrents containing several files, expand the top-level row and select the individual child file before using Download or Copy URL.

Delete safely

Sylve refuses to delete an item while it is pending or processing. It also refuses deletion when a VM currently references that download, or when it cannot safely prove which managed files belong to the entry. These checks prevent a cleanup action from removing media that a VM still needs or unrelated files on the node.

Remove or replace any VM media attachment that references the item, wait for active work to finish, then try deletion again. Eject alone is not enough: it keeps the attachment and its reference to the download. To remove that reference, shut the VM off and use Delete on its media row in Storage. Deleting the media attachment leaves the source in Downloader until you delete it here. If the table reports retained paths after an unsuccessful cleanup, review those paths on the node before making manual changes.

Use a downloaded image in a VM

Wait for the download and any processing to reach Done, then shut the VM off and open Storage → Add storage:

• Choose Create disk from image for a writable guest disk. Sylve copies or converts the selected disk image into a separate managed ZFS volume or RAW file. The Downloader source remains unchanged. You do not need to enable Auto-convert to RAW in Downloader solely for this workflow.
• Choose Attach read-only media for an installer ISO. This references the download directly without creating a writable copy.

Extract compressed disk images before selecting them. Torrent downloads do not support automatic extraction, so extract those images externally and add the extracted file by its node-local path. Images that depend on external backing files must be made self-contained first. Installer ISOs cannot be used with Create disk from image.

For a writable disk, leave the optional target size empty to use the image’s virtual capacity, or choose a larger capacity. A larger disk does not automatically expand partitions or filesystems inside the guest. See VM Storage for the complete workflow.

Practical examples

Goal	Source and settings
Boot a VM from a FreeBSD or Linux installer	Use the publisher’s HTTPS ISO URL. Choose Uncategorized and leave processing disabled.
Create a writable disk from an appliance image	Add the disk image as Uncategorized, extracting it if compressed. Once it is Done, use Create disk from image in VM Storage. Conversion to RAW in Downloader is not required.
Create a VM from a cloud image with first-boot configuration	Add the publisher’s cloud image by HTTPS or upload it. Choose Cloud-Init and enable Auto-convert to RAW if the source is not already RAW.
Create a jail from a base archive	Add the base archive by HTTPS, local path, or upload. Choose Base / RootFS. Extraction is enabled automatically.
Reuse an image already on the node	Enter the file’s absolute node path, choose the appropriate type, and optionally process it. Sylve copies it into its managed library.