Read-only file system error in /media/fat/cifs

[svenTeum]

I am getting the error below every time I run update_all.sh. Because of that, every time the complete file check is executed and it takes a long time.

/media/fat/cifs is where an external NAS unit is mounted via cifs_mount.sh script. No other errors are shown and the "games" folder that is in /media/fat/cifs is checked OK. I also checked via mister Linux console that I can create a file in the /media/fat/cifs path.

I included the full log as an attachment.

DEBUG| Saving json "/media/fat/Scripts/.config/downloader/downloader.json"
DEBUG| Saving json "/media/fat/Scripts/.config/downloader/downloader_sigs.json"
DEBUG| Saving json "/media/fat/cifs/.downloader_db.json"
Traceback (most recent call last):
  File "/tmp/ua_downloader_dd.pyz/downloader/main.py", line 17, in main
    exit_code = execute_full_run(
  File "/tmp/ua_downloader_dd.pyz/downloader/main.py", line 61, in execute_full_run
    exit_code = runner.full_run()
  File "/tmp/ua_downloader_dd.pyz/downloader/full_run_service.py", line 57, in full_run
    result = self._full_run_impl()
  File "/tmp/ua_downloader_dd.pyz/downloader/full_run_service.py", line 106, in _full_run_impl
    save_store_err = self._local_repository.save_store(install_box.local_store())
  File "/tmp/ua_downloader_dd.pyz/downloader/local_repository.py", line 225, in save_store
    self._file_system.save_json(store, os.path.join(drive, FILE_downloader_external_storage))
  File "/tmp/ua_downloader_dd.pyz/downloader/file_system.py", line 413, in save_json
    with open(full_path, 'wb') as f:
OSError: [Errno 30] Read-only file system: '/media/fat/cifs/.downloader_db.json'

Thank you.

update_all.log.tar.gz

[theypsilon]

Hey! It seems to be a problem in how that partition is mounted for you. The real fix here would be addressing that in the OS level, which is beyond the scope of the updaters. You could try to see how is the partition tab configured and check there why the OS reported that partition to be read-only.

From the updaters side, I'll do a better error handling there with a more descriptive message. That will save you some time, but the underlying problem will still be there, and the updater won't be able to sync file removals from the databases since no accounting is being stored in that partition.

In case you address the OS issue, it would be helpful to post here here how did you fix it, so that other users can learn from it.

[svenTeum]

Thanks for looking at it. It seems that the unit was being (sometimes) mounted as read-only, but I'm not sure about the cause. I use samba to create the NAS share and, as I said, cifs_mount script to mount it in Mister. To avoid launching the update_all before knowing if the unit is ok, I added an auxiliary script to my setup that just tests if the unit (/media/fat/cifs) is writable. I thought the update_all script may be related since sometimes the unit was writable but suddenly it became read-only, but it is clearly not. I will add more information in case I find the cause.

About the script, maybe it would be nice to check that the unit is OK before the long process to avoid failing at the end, which causes the process to repeat and fail again. But I understand that this is a very particular case.

You can close the issue. Thank you.