We use cookies to understand how the site is used and to improve it. You can accept them, or carry on with them switched off.
Also known as: out of disk space, no space left
Downloads pause or fail when the destination volume runs out of space, and the client often reports only a generic write error. A Usenet workflow needs room for the incomplete folder, the extracted output, and the temporary files PAR2 repair creates, which together can approach three times the release size. Keeping incomplete and completed folders on the same small volume makes this far more likely.
The arithmetic behind this catches people out because a release occupies its own size more than once, and the peak arrives during post-processing rather than at the moment the last segment lands. A job that fitted comfortably when it was queued can exhaust the volume during repair, after every byte has already been paid for once. Fault here is local without qualification: no provider setting and no indexer affects it.
Three claims on the volume overlap in time.
.par2 recovery volumes, which for a typical posting add somewhere in the region of five to fifteen percent on top of the set.Repair adds temporary files of its own while it reconstructs damaged blocks. Two and a half to three times the release size is a safe planning figure, and the requirement is higher where the completed folder sits on a different filesystem from the incomplete folder, because the final move is then a copy and both copies exist at once.
A volume can report available gigabytes and still refuse writes. Snapshots on ZFS or Btrfs pin deleted data until they expire. Recycle bins retain cleared downloads. Reserved blocks on ext4 are unavailable to an unprivileged process. On a container host, image layers and log files grow somewhere nobody is watching. Thin-provisioned storage can be exhausted at the pool level while the guest filesystem still believes it has room, and a per-user quota produces an identical failure on a volume with plenty to spare.
The client sees a failed write and reports what the operating system handed it, which is often a generic I/O error. The same shortage surfaces as an unpack that fails partway, as a repair that aborts in a way easily mistaken for insufficient recovery blocks, or as a queue that simply pauses itself. Checking free space is worth doing before any of those are investigated as separate problems, since one shortage produces all of them.
Freeing space and resuming is usually enough, and completed segments in the incomplete folder survive. Partial output from an interrupted extraction should be removed first, since some clients will not overwrite it and others will treat it as finished. The durable fix is separating the roles: the incomplete folder on a volume sized for working space, the completed folder on the same filesystem as the final library so that the move is a rename rather than a copy, and a minimum free-space threshold configured in the client so that it pauses rather than fails.
Browse all terms in Errors & Troubleshooting, or see the full Usenet glossary.