2024-11-07T14:09:31.696Z

<Igor Golikov> Hi folks, I am hitting
`Link Error: RADOS library not found`
when trying to build latest ceph main on my machine.
Did anyone see it as well?

2024-11-07T15:03:34.791Z

<Yehuda Sadeh-Weinraub> 1. DeleteObject (with versionId) should succeed as long as snapshot id is not specified, hence dealing with the current state. If this object is part of a snapshot then its index entry should be updated to reflect at what snapshot id it was removed
2. An object entry in the index would have (at least) two new fields: its own snapshot id, and "skip_lower_snap_obj". When creating a new object instance, it would then look at the next object instance in the index and: if the next object has a lower snapshot id it would set skip_lower_snap_obj to the next object key, if it's not lower it would copy its skip_lower_snap_obj field.

2024-11-07T19:11:56.484Z

<Ivveh> hello there, anyone from mgr/dashboard that can shed a little light on a thingie? im trying to solve a bug where modifying exports via the dashboard incorrectly restarts the nfs-ganesha deployment because `fsal['user_id']` is empty in dashboard/controllers/nfs.py , this later in nfs/export.py becomes None so a comparison fails and then incorrectly sets `need_nfs_service_restart` to true. I see that the documentation helper for the `CREATE_EXPORT_SCHEMA` doesnt contain `user_id` in `fsal` but its there in `EXPORT_SCHEMA`, however when trying to do a change the key is not there, im guessing this is intended.. we could always add:
```if new_fsal.user_id is None:
    new_fsal.user_id = old_fsal.user_id```
in nfs.py (as it correctly works when modifying an export via API) but other things might break.. but at least this doesn't restart ganesha 🤷

2024-11-07T19:14:19.226Z

<Ivveh> hello there, anyone from mgr/dashboard that can shed a little light on a thingie? im trying to solve a bug where modifying exports via the dashboard incorrectly restarts the nfs-ganesha deployment because `fsal['user_id']` is empty in dashboard/controllers/nfs.py , this later in nfs/export.py becomes None so a comparison fails and then incorrectly sets `need_nfs_service_restart` to true. I see that the documentation helper for the `CREATE_EXPORT_SCHEMA` doesnt contain `user_id` in `fsal` but its there in `EXPORT_SCHEMA`, however when trying to do a change the key is not there, im guessing this is intended.. we could always add:
```if new_fsal.user_id is None:
    new_fsal.user_id = old_fsal.user_id```
in nfs.py (as it correctly works when modifying an export via API) but other things might break.. but at least this doesn't restart ganesha when doing it in the dashboard🤷🏻

2024-11-07T19:26:34.245Z

<Zac Dover> This is a big docs project (and maybe not really a docs project entirely, but some kind of project that works to standardize the whole ceph-config situation), but I have created a docs tracker here: <https://tracker.ceph.com/issues/68874>

This doesn't mean that we're going to have a comprehensive guide to Ceph config by Christmas, but it's a first step toward improving the situation (through the medium of docs).

2024-11-07T19:31:31.113Z

<Ivveh> and it seems that if its an rgw export it will always restart, no matter what

2024-11-07T20:06:41.346Z

<Ken Carlile> of course, thank you!