# File system repository returning error

**URL:** <https://discuss.elastic.co/t/file-system-repository-returning-error/379330>\
**Category:** Elasticsearch\
**Tags:** snapshot-and-restore\
**Created:** [June 19, 2025, 2:24pm UTC](https://discuss.elastic.co/t/file-system-repository-returning-error/379330 "2025-06-19T14:24:39Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![JSElasticDiscuss](https://avatars.discourse-cdn.com/v4/letter/j/3bc359/32.png) [@JSElasticDiscuss](https://discuss.elastic.co/u/JSElasticDiscuss)\
**Post date:** [June 19, 2025, 2:24pm UTC](https://discuss.elastic.co/t/file-system-repository-returning-error/379330/1 "2025-06-19T14:24:39Z")

</div>

Hello,

I have confirmed a valid path in my `path.repo` setting, but now when I do a 'verify repository' it gives me an error:

```auto
{
  "name": "ResponseError",
  "message": "repository_verification_exception\n\tCaused by:\n\t\tdirectory_not_empty_exception: /mnt/elastic-share/es-backups/tests-QVpHYVK-S26hF3FqOI6dxg\n\tRoot causes:\n\t\trepository_verification_exception: [Before-818upgrade-6_19_2025] cannot delete test data at "
}

```

And the verification status remains 'Not Connected'  
The `'tests-<anything>'` continues to appear in there without getting deleted. I checked permissions but it's wide open for `elasticsearch` user. I also confirmed that after deleting the tests item that appears in there, if I click the verify repository again, it fails with the same error and a new `'tests-<anything>'` has appeared.

We're working on performing an upgrade to 9.x, so have to upgrade to 8.18 first, which is where this started appearing.

---

<div class="post-metadata">

**Author:** ![DavidTurner](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/davidturner/32/22453_2.png) [@DavidTurner](https://discuss.elastic.co/u/DavidTurner)\
**Post date:** [June 23, 2025, 8:55am UTC](https://discuss.elastic.co/t/file-system-repository-returning-error/379330/2 "2025-06-23T08:55:39Z")

</div>

This means your repository at `/mnt/elastic-share/es-backups/` is not behaving like a proper filesystem (at least, not one with Elasticsearch-exclusive access). Elasticsearch created a handful of files in `/mnt/elastic-share/es-backups/tests-QVpHYVK-S26hF3FqOI6dxg/` and then deleted them, but either the deletion did not work or something else also created a file in this directory at the same time. Either way, the directory wasn't empty afterwards so couldn't be deleted, which indicates that this filesystem is not safe to use as a snapshot repository.

---

<div class="post-metadata">

**Author:** ![JSElasticDiscuss](https://avatars.discourse-cdn.com/v4/letter/j/3bc359/32.png) [@JSElasticDiscuss](https://discuss.elastic.co/u/JSElasticDiscuss)\
**Post date:** [June 26, 2025, 1:58pm UTC](https://discuss.elastic.co/t/file-system-repository-returning-error/379330/3 "2025-06-26T13:58:15Z")

</div>

Thanks @DavidTurner -- I managed to get the backup configured using the top folder /mnt/elastic-share as the location. This worked, but I am curious if we have the backup folder configured incorrectly?  
I didn't see any documentation describing who the owner of that repo should be or the permissions it should have. Do you have that information handy?  
Currently it's configured as  
o/g: Root perm: 644

---

<div class="post-metadata">

**Author:** ![DavidTurner](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/davidturner/32/22453_2.png) [@DavidTurner](https://discuss.elastic.co/u/DavidTurner)\
**Post date:** [June 27, 2025, 8:57am UTC](https://discuss.elastic.co/t/file-system-repository-returning-error/379330/4 "2025-06-27T08:57:42Z")

</div>

No special configuration required here. ES isn't doing any magic, it's just creating/reading/writing files using the standard filesystem syscalls, like any other userspace process.