Assistance with mount_shares_locally to establish correct symlink/pool data deletion behaviour

[astrom0]

Please help me with:
https://github.com/gboudreau/Greyhole/wiki/MountSharesLocally
to achieve the objective of deleting a symlink from the LZ and have it reliably reflected down to the gh data pool, so that the linked file/directory is deleted.

Item 1. Shares of Shares?
I got mixed up and added data over to the landing zone samba shares directly:
localhost:8012/?page=id_l1_smbshares_tab
Data added this way have been correctly managed by gh using fsck and balance, as far as I can tell.
I missed or misunderstood the MountSharesLocally wiki advice. The penny has dropped that I should be interacting with samba shares of the greyhole LZ samba shares, as per comments given in: #271
If adding files directly to the LZ messes anything up, please advise and offer suggestions of how to fix issues that this may have caused.

Item 2. MountSharesLocally Directory Ownership
Next point, after following the steps in https://github.com/gboudreau/Greyhole/wiki/MountSharesLocally, I can't copy files over to the LZ samba shares that I've got set up.

When I check folder ownership by running ~$ ls -l /mnt/samba/ all the folders are owned by root, e.g.:
drwxr-xr-x 2 root root 4096 Mar 16 08:59 LZ_Folder

I've gone ahead and changed the ownership of this directory to suit my user_name.

Is this a step that https://github.com/gboudreau/Greyhole/wiki/MountSharesLocally overlooks?
Am I messing anything up by taking ownership?
If I don't take ownership, how is writing to /mnt/samba/ meant to work?

Item 3. No Data in /mnt/samba/
After all this, I see the LZ_Folder show up in /mnt/samba/ but there is no data revealing within the folder itself.

I've rebooted the machine for good measure.

I spotted this piece of dialogue:
#305
So, I updated /etc/samba/smb.conf to include:

follow symlinks = yes
wide links = yes

under [LZ_Folder] instead of having this detail specified under [global]

Restarting the service &/or rebooting doesn't change behaviour of /mnt/samba/

I also spotted this piece of history: #297
but I'm on greyhole version 0.15.25, so that user's problem shouldn't be relevant anymore.

So, what should I do to establish proper behaviour from greyhole?

As noted at the top, my objective here is to be able to delete a symlink from the LZ and have it reliably reflected down to the gh data pool so that the linked file(s)/directory(ies) is(are) deleted.

Thanks for the help.

[gboudreau]

1. greyhole --fsck will do the same thing the Greyhole daemon would have done, had you added those files using a Samba shares. What it won't have handled properly are renames (or moving files & folders) and deletions. If you did those, you'll have to clean up manually.

2. You say you followed the steps in the MountSharesLocally article, but in point 3, you say that the folders in /mnt/samba/ are empty. So you have a problem right there. You need your local shares to be successfully mounted before you can interact with them, so you need to fix your point 3 first, then the files & folders permissions should not be a problem.

3. You need to look into why the mounts seem to fail. Look into the init.d logs. Once you run sudo /etc/init.d/mount_shares_locally restart, if something it not right, it should complain, and tell you to use journalctl or something to look into the error.
   You can check if the local Samba mounts were correctly mounted using the mount command.
   Or go into one of them, and use df . to check for free space. It will tell you (look in the Filesystem value) if the device that contains the current folder is a hard drive, or a samba share:

gb@fileserver5:/ $  df .
Filesystem     1K-blocks     Used Available Use% Mounted on
/dev/md0       152595792 64917904  79853616  45% /

gb@fileserver5:/ $ cd /mnt/samba/Videos/

gb@fileserver5:/mnt/samba/Videos $  df .
Filesystem                1K-blocks        Used  Available Use% Mounted on
//192.168.155.10/Videos 79032669996 76018781360 3013888636  97% /mnt/samba/Videos

To get further help, use this script to provide more info about your installation: https://github.com/gboudreau/Greyhole/wiki/GetHelp#gather-system-information

[astrom0]

Thanks for that tip.

Based on this response:

astrom0@server:/mnt/samba$ ls
'print$'  'u ts653'   u_ts653  'w ts653'   w_ts653  'x ts653'   x_ts653  'z ts653'   z_ts653
astrom0@server:/mnt/samba$ cd /mnt/samba/u_ts653
astrom0@server:/mnt/samba/u_ts653$ df .
Filesystem                        1K-blocks     Used Available Use% Mounted on
/dev/mapper/ubuntu--vg-ubuntu--lv 111640908 84119728  21803984  80% /

and your advice, there is clearly something off.

I've got an fsck running and need to head off, so I'll follow up as you suggest and report back.

From memory, I don't believe I saw any complaints when I ran sudo /etc/init.d/mount_shares_locally restart previously but I'll double check on that point to make sure.

[astrom0]

Back again, for clarity/confirmation, I stopped the mount_shares_locally script and went through the whole protocol from https://github.com/gboudreau/Greyhole/wiki/MountSharesLocally
and deleted all the sub-directories in /mnt/samba/

[As a double check, please confirm that the correct syntax for the mount_shares_locally script on line 12 is to retain the double quotes around your username]

The bump in the process of applying https://github.com/gboudreau/Greyhole/wiki/MountSharesLocally was at step 5:

~$ sudo /etc/init.d/mount_shares_locally start
Starting mount_shares_locally (via systemctl): mount_shares_locally.serviceWarning: The unit file, source configuration file or drop-ins of mount_shares_locally.service changed on disk. Run 'systemctl daemon-reload' to reload units.

So, I ran systemctl daemon-reload and there was no further complaint when I gave the command sudo /etc/init.d/mount_shares_locally start

However, system behaviour was the same as previously described.

I applied https://github.com/gboudreau/Greyhole/wiki/GetHelp#gather-system-information but ix.io is not currently functional:

~$ curl -Ls https://bit.ly/gh-infos | sudo sh
Here's the URL you will need to give the to person who's helping you:
<!doctype html>
<html lang=en>
<head>
    <meta charset=utf-8>
    <title>&nbsp;</title>
</head>
<style>
    body {
        font-family: monospace;
        margin: 2em;
    }
</style>
<body>
    <p>ix.io is taking a break &#127867;</p>
    <img src="/underconstruction.gif" width="200px">
</body>
</html>

So, where to from here? What other way can I pass on the relevant logs?

[gboudreau]

So, where to from here? What other way can I pass on the relevant logs?

The relevant logs are found in /tmp/greyhole_sysinfo. Just cat that file afterwards, and paste it in a new comment in this issue.