Fix it now
These are download manager codes, raised after a scan has already succeeded. Microsoft publishes 0x80246001 as a requested file with no URL, and 0x80246005 as a download attempted with no network connection available. Metadata worked and content did not, which narrows the search a great deal.
sc query bits
net start bits
net stop wuauserv
ren C:\Windows\SoftwareDistribution SoftwareDistribution.old
net start wuauserv
- Check the connection in the machine’s own terms, not your browser session, and confirm the network is not set as metered under Settings, Network and internet.
- Clear stuck transfers from an elevated PowerShell prompt:
Get-BitsTransfer -AllUsersto see them, thenGet-BitsTransfer -AllUsers | Remove-BitsTransferif jobs are sitting in an error or suspended state. - If the device is managed by WSUS, check on the server that content for the approved update has actually been downloaded rather than only approved.
- Rescan and retry the download, watching whether the percentage moves at all.
Removing BITS jobs discards partial downloads. Windows starts them again from the beginning, which is inconvenient on a slow line and harmless otherwise.
If the download now runs, stop here. Below is why metadata and content travel separately, and what each of the four codes is actually reporting.
Why it happens
A scan and a download are two different journeys. The scan returns metadata: which updates exist, which apply to this device, and where their files can be obtained. The download is a separate operation handed to BITS and Delivery Optimization, which fetch the payload from whatever sources that metadata named. They travel over different paths and fail independently, which is exactly why a machine can list available updates perfectly and then refuse to download any of them.
Microsoft’s published descriptions make the split obvious. 0x80246001 is WU_E_DM_URLNOTAVAILABLE: a download manager operation could not be completed because the requested file does not have a URL. There is no network fault in that sentence at all; the client was handed a file reference with nowhere to get it from. 0x80246005 is WU_E_DM_NONETWORK: the operation could not be completed because the network connection was unavailable. Those are opposite problems and they need opposite investigations.
The two supporting codes are narrower. 0x80246003 is WU_E_DM_UNKNOWNALGORITHM, meaning the file metadata requested a hash algorithm the download manager does not recognise, which points at a client too old for the content it is being offered. 0x80246004 is WU_E_DM_NEEDDOWNLOADREQUEST: an operation could not be completed because a download request is required from the download handler, which is what a caller gets when it asks about a transfer it never queued.
WSUS approved the update but holds no content for it
You have this one if 0x80246001 on managed clients only, while unmanaged machines on the same network download normally.
- On the WSUS server, check whether update files are stored locally or whether clients are meant to download from Microsoft directly.
- If files are stored locally, confirm content synchronisation has completed and is not stalled behind a large backlog.
- Retry the download from a client once the server reports the content present.
This is the single most common cause of 0x80246001 in managed estates, and it is a server-side condition. Nothing you run on the client will produce a URL the server never supplied.
There is no usable network in the machine context
You have this one if 0x80246005, on a captive portal, a split-tunnel VPN, or a connection marked as metered.
- Complete any captive portal sign-in before starting the download.
- Turn off the metered setting for that network, or allow downloads over metered connections in Settings, Windows Update, Advanced options.
- On a VPN, confirm the update endpoints are routed rather than dropped by the tunnel policy.
- Retry on a stable connection, ideally wired for anything the size of a feature update.
BITS is stuck or has been disabled
You have this one if Downloads sit at zero per cent, Get-BitsTransfer -AllUsers lists old jobs, or the service will not start.
- Remove stale jobs with
Get-BitsTransfer -AllUsers | Remove-BitsTransfer. The -AllUsers switch needs administrative credentials. - Put the service back to its normal start type with
sc config bits start= delayed-auto, thennet start bits. - Read the BITS entries in Event Viewer under Applications and Services Logs for repeated failures.
Delivery Optimization has been left with nowhere to fetch from
You have this one if Downloads fail on a network configured for peer sources, or on an isolated site where the peers do not exist.
- Check Delivery Optimization under Settings, Windows Update, Advanced options.
- In a managed environment, read the download mode policy. Mode 1 is LAN and is the default, mode 2 is Group, mode 3 adds internet peers, mode 0 is HTTP only and mode 99 is Simple for offline environments.
- Mode 100, Bypass, is deprecated in Windows 11; Microsoft’s guidance is to use 0 or 99 instead.
- Retry the download after changing the mode.
Full reference
The four codes, exactly as published
| Code | Published name | Published description |
|---|---|---|
0x80246001 |
WU_E_DM_URLNOTAVAILABLE | The requested file does not have a URL |
0x80246003 |
WU_E_DM_UNKNOWNALGORITHM | The file metadata requested an unrecognised hash algorithm |
0x80246004 |
WU_E_DM_NEEDDOWNLOADREQUEST | A download request is required from the download handler |
0x80246005 |
WU_E_DM_NONETWORK | The network connection was unavailable |
Delivery Optimization download modes
Where peer sourcing is configured, the mode decides whether the client is allowed to reach the original source at all, and a restrictive mode on a network with no peers is a slow way to discover that.
| Mode | Name | What it allows |
|---|---|---|
| 0 | HTTP Only | No peer caching. Downloads come from the original source or a Microsoft Connected Cache server |
| 1 | LAN | The default. Peer sharing with devices on the same network |
| 2 | Group | Peering across internal subnets between devices in the same group |
| 3 | Internet | Adds internet peer sources |
| 99 | Simple | Delivery Optimization cloud services disabled, HTTP only, for offline environments |
| 100 | Bypass | Deprecated in Windows 11. Microsoft recommends mode 0 or 99 instead |
Working out which half of the journey failed
- Open Settings, Windows Update and confirm updates are listed. If they are, metadata is fine and this is genuinely a content problem.
- Watch the percentage. Nothing at all points at BITS or at a missing URL; a stall partway through points at the connection or the source.
- Run
Get-BitsTransfer -AllUserswhile the download is supposed to be running. An absent job and a failed job mean different things. - On a managed client, compare against an unmanaged machine on the same network. A difference there puts the fault on the update server rather than the network.
- As a last resort, take the package from the Microsoft Update Catalog matching your build, edition and architecture exactly, and install it locally while you fix the source.
What not to spend time on
Two things absorb a lot of effort here and rarely help. Resetting the whole update client is a reasonable move once, and pointless the second time, because a missing URL is a property of the metadata rather than of the client’s state. And repairing the component store with DISM addresses a different failure entirely: these codes are raised before any payload is applied, so a healthy or unhealthy component store is not what decided the outcome.
Logs worth reading
| Source | How to get it | What it tells you |
|---|---|---|
| Windows Update trace | Get-WindowsUpdateLog |
Which files the client asked for and what it was given for each |
| BITS events | Event Viewer, Applications and Services Logs | Repeated transfer failures and the reason each job stopped |
| Delivery Optimization | Settings, Windows Update, Advanced options | Where content actually came from on recent downloads |
Every code this article covers
| Code | What it points at | Source |
|---|---|---|
0x80246001 |
WU_E_DM_URLNOTAVAILABLE. A download manager operation could not be completed because the requested file does not have a URL | Microsoft Learn |
0x80246003 |
WU_E_DM_UNKNOWNALGORITHM. A download manager operation could not be completed because the file metadata requested an unrecognised hash algorithm | Microsoft Learn |
0x80246005 |
WU_E_DM_NONETWORK. A download manager operation could not be completed because the network connection was unavailable | Microsoft Learn |
0x80246004 |
WU_E_DM_NEEDDOWNLOADREQUEST. An operation could not be completed because a download request is required from the download handler | Microsoft Learn |
Confirm the fix worked
- The download that previously failed reaches 100 per cent and the update installs.
- It appears in Settings, Windows Update, Update history.
Get-BitsTransfer -AllUsersshows no jobs left in an error or suspended state.- On a managed client, the update server reports the device as compliant after the next detection cycle.
Questions people ask about this
Does fixing this involve buying anything?
No. These are connectivity and server content problems, and every step here is free. Updates themselves cost nothing on a supported, licensed installation.
Why do updates list correctly but refuse to download?
Because listing and downloading use different mechanisms. The scan needs only metadata. The download needs a reachable source for the actual files, and that is the part that failed.
Is turning off the metered connection setting risky?
Only to your data allowance. If the connection really is metered, use a different network for large downloads rather than leaving the setting off permanently.
Can I just download the update by hand?
Yes. Take it from the Microsoft Update Catalog, matching your build, edition and architecture exactly, and install it locally. It is a reasonable stopgap while you fix the underlying source problem, but it is not a substitute for fixing it.
What does 0x80246004 tell me?
That something asked about a download that was never requested. In practice it means a script or management tool queried transfer progress before queueing the transfer, so the fix is in the calling code rather than on the machine.
