# Apparent compression header error on 0.9.7

**URL:** <https://kopia.discourse.group/t/apparent-compression-header-error-on-0-9-7/822>\
**Category:** Support\
**Created:** [December 20, 2021, 3:50pm UTC](https://kopia.discourse.group/t/apparent-compression-header-error-on-0-9-7/822 "2021-12-20T15:50:42Z")\
**Posts on this page:** 1\
**Showing post:** 7

<div class="post-metadata">

**Author:** ![nickersonm](https://yyz2.discourse-cdn.com/free1/user_avatar/kopia.discourse.group/nickersonm/32/162_2.png) [@nickersonm](https://kopia.discourse.group/u/nickersonm)\
**Post date:** [January 1, 2022, 8:27pm UTC](https://kopia.discourse.group/t/apparent-compression-header-error-on-0-9-7/822/7 "2022-01-01T20:27:17Z")

</div>

Update: the two referenced blobs, p5a59d… and p5a596… are both zero bytes. Is there any way to either recover these or determine which snapshots they belong to? If they are old versions of files, I would prefer to drop that snapshot rather than restart the entire backup from scratch.

Edit: I attempted to use `kopia content delete` to remove the contents I didn’t have valid blobs for, accepting the data loss. However, then `kopia maintenance run --full` completed with:

`ERROR snapshot GC failure: error running snapshot gc: error iterating contents: Could not undelete referenced content: {a��+��♣�↓��↓�� 0315933df5321eb1a7d729b3ec26fdcb 0xc00037a480}: unable to get content data and info: error getting cached content: BLOB not found`

Following [this thread](https://kopia.discourse.group/t/is-one-bit-flip-enough-to-crash-a-kopia-repo-how-to-recover-from-bit-rot/316/4), I re-snapshotted everything with `--force-hash=100`. The full maintenance still did not complete successfully - it appears I am unlucky and an altered file was corrupted, so unless [this issue](https://github.com/kopia/kopia/issues/799) is implemented I will continue having GC errors.

---

_[View the full topic](https://kopia.discourse.group/t/apparent-compression-header-error-on-0-9-7/822)._