[Plugin] CA Fix Common Problems


Recommended Posts

The powerdown script built into v6.2 is vastly improved over the v6.1 script and does a good job at shutting down the server without hanging up in an endless loop.  The powerdown plugin offers some advantages such as K and S script processing on shutdown and start up (although these situations can be handled as well now with the user_scripts plugin), more extensive shutdown logging, and archiving of logs.

 

Where I am going with this is I think you should suggest the powerdown script as an option (or say nothing at all) and not show it as a must have plugin.  It's not as necessary now as on earlier versions.

 

As LT improves the shutdown process, I can see the powerdown script no longer necessary on later versions and I would like to eventually deprecate the plugin.  I am thinking that I'll block its install on versions later than v6.2.  I have been uncomfortable with taking over the unRAID shutdown process and I doubt I'll want to invest more time in the plugin in the future.

ok.  This what I'll do then.

 

I'll downgrade the level of powerdown not being installed from an error to a warning (although it is still very commonly told to users to install it whenever there are any sort of issues at all)

 

If/when you block its installation on 6.2 then I'll adjust FCP to not complain at all if the user is running 6.2 (and leave the warning on 6.1)

 

Additionally, when you block its installation, adjust the plugin xml with the <MaxVer> on it.  That way CA won't offer it to people at all on 6.2, and if FCP finds it installed (installed prior to you deprecating it and after upgrading to the blocked version of unRaid) then it will generate an error.

 

Powerdown will not install on v6.2 any more.  I've added the <MaxVer> tag to 6.1 so CA will no longer show it as a plugin to install.

 

It is no longer necessary.

Sure... let me know right after I release an update to adjust for the fact that the RC candidates were 6.2.0-RCx and final is 6.2 (technically, 6.2.0-RCx is a later release than 6.2)

Just read it a bit closer.  Did you set it to 6.1 or 6.1.9  Big difference there, as both CA and FCP take into account point releases (along with Betas and RCs)

 

6.1.

Should probably change it then

 

Done.  Set to 6.1.9.  As I understand, that is the max version CA will allow and FCP will check.

Correct.  And CA is already flagging it as incompatible on my 6.2 system and my 6.1.9 system its ok.

 

Just have to reconfigure FCP and we're good to go.

 

Thank you.

Link to comment

The /sbin/powerdown will still exist - it will just be the standard Limetech version rather than the one provided by the Powerdown plugin.

 

server didn't power down last night ;(

 

from cli

root@MovieRaid:~# /sbin/powerdown

-bash: /sbin/powerdown: No such file or directory

 

 

/usr/local/sbin/powerdown

 

or just

 

powerdown

Link to comment

The /sbin/powerdown will still exist - it will just be the standard Limetech version rather than the one provided by the Powerdown plugin.

 

server didn't power down last night ;(

 

from cli

root@MovieRaid:~# /sbin/powerdown

-bash: /sbin/powerdown: No such file or directory

 

 

/usr/local/sbin/powerdown

 

or just

 

powerdown

 

powerdown or /usr/local/sbin/powerdown does NOTHING ....

 

going to install the package again after i find this restriction and delete the restriction

 

Link to comment

The /sbin/powerdown will still exist - it will just be the standard Limetech version rather than the one provided by the Powerdown plugin.

 

server didn't power down last night ;(

 

from cli

root@MovieRaid:~# /sbin/powerdown

-bash: /sbin/powerdown: No such file or directory

 

 

/usr/local/sbin/powerdown

 

or just

 

powerdown

 

powerdown or /usr/local/sbin/powerdown does NOTHING ....

 

going to install the package again after i find this restriction and delete the restriction

 

At the cli:

root@MediaServer:~# whereis powerdown
powerdown: /usr/local/sbin/powerdown
root@MediaServer:~#

 

The powerdown script is there.  I suspect that there is an issue in your server preventing the shutdown that the plugin may have helped with.  The plugin was deprecated for several reasons.  The main reason is that it was not really necessary because the builtin powerdown is more robust.  Another is that v6.2 presented some challenges that I didn't overcome so it was not ready for v6.2 release.

 

Rather than getting all dramatic and overcoming the restriction to install a deprecated plugin, why don't you spend a few minutes and find out the problem with the built in powerdown.  At the command line type powerdown and then troubleshoot the issue(s).

 

EDIT: Take a look at this post: http://lime-technology.com/forum/index.php?topic=31735.570#lastPost

 

We are back to the reason the powerdown plugin was created in the first place.  The webgui shutdown gets stuck in a loop waiting for an event that will never happen because the powerdown script is relying on the webgui to perform a shutdown.  I think it's time for LT to finally solve this nagging problem.  I'm going to post a defect report.

Link to comment

well installing the powerdown plugin solves the issue again

 

array shut down nicely last night

 

i don't mind troubleshooting but 2 times i tried powerdown without plugin and really "NOTHING" happens

no umounts not nothing

it is like typing clear in terminal but at least clear cleans up your terminal

powerdown did nothing lol

 

 

Link to comment

Last week I moved my unRAID over to an Active Directory setup. I put everything back to default nobody/users, removed all permissions from world, and then applied AD permissions using AD groups. Today when my weekly extended tests ran, I discovered that the plugin really doesn't like AD permissions... :) 

 

"The following files / folders may not be accessible to the users allowed via each Share's SMB settings."

 

This section then listed every single file I have, lol!

Link to comment

Would it be possible for this check to provide info on which disks have the files? It's not too hard to find on your own, but since the plugin already knows......

 

The following files exist within the same folder on more than one disk.  This duplicated file means that only the version on the lowest numbered disk will be readable, and the others are only going to confuse unRaid and take up excess space:
Link to comment

Last week I moved my unRAID over to an Active Directory setup. I put everything back to default nobody/users, removed all permissions from world, and then applied AD permissions using AD groups. Today when my weekly extended tests ran, I discovered that the plugin really doesn't like AD permissions... :) 

 

"The following files / folders may not be accessible to the users allowed via each Share's SMB settings."

 

This section then listed every single file I have, lol!

lol  Must have been quite the email notification...  Can't test AD integration, so just ignore the error, but I'll look at limiting the results

 

Would it be possible for this check to provide info on which disks have the files? It's not too hard to find on your own, but since the plugin already knows......

 

The following files exist within the same folder on more than one disk.  This duplicated file means that only the version on the lowest numbered disk will be readable, and the others are only going to confuse unRaid and take up excess space:

NP
Link to comment

Not sure where to go with this, but I ran extended tests and to my surprise I have a lot of duplicate files.  I am addressing the issues and re-starting the test to see more, but the results are cumulative and don't clear between tests.  Is there a way to clear the previous results so I can start a new test fresh?

Link to comment

Not sure where to go with this, but I ran extended tests and to my surprise I have a lot of duplicate files.  I am addressing the issues and re-starting the test to see more, but the results are cumulative and don't clear between tests.  Is there a way to clear the previous results so I can start a new test fresh?

I had them as cumulative so that you could go back and reference the old results, but since I always have to force problems to happen, I've never really given it a second thought to re-running it a few times in succession.

 

Painless for me to delete the prior results.  Will add it in to the changes for the posts above.  Look for it tomorrow ish.

 

But in the meantime, you can delete /tmp/fix.common.problems/extendedLog

Link to comment

Not sure where to go with this, but I ran extended tests and to my surprise I have a lot of duplicate files.  I am addressing the issues and re-starting the test to see more, but the results are cumulative and don't clear between tests.  Is there a way to clear the previous results so I can start a new test fresh?

I had them as cumulative so that you could go back and reference the old results, but since I always have to force problems to happen, I've never really given it a second thought to re-running it a few times in succession.

 

Painless for me to delete the prior results.  Will add it in to the changes for the posts above.  Look for it tomorrow ish.

 

But in the meantime, you can delete /tmp/fix.common.problems/extendedLog

 

Thanks.  That would be much better to start fresh each time.  I couldn't believe the issues I had with duplicate files.  It was pages and pages of them.  I don't know where that came from.

 

Suggestion:  You are looking in the appdata folder for characters that Windows doesn't like.  In my case I think FCP was complaining about the ':' character.  That folder is hidden and used by Dockers that are Linux based.  That folder shouldn't even be browsed or manipulated by Windows.  Can that one be skipped for that check?

Link to comment

Suggestion:  You are looking in the appdata folder for characters that Windows doesn't like.  In my case I think FCP was complaining about the ':' character.  That folder is hidden and used by Dockers that are Linux based.  That folder shouldn't even be browsed or manipulated by Windows.  Can that one be skipped for that check?

There's just no satisfying you is there  ;)  Curious though what container its from...
Link to comment

Suggestion:  You are looking in the appdata folder for characters that Windows doesn't like.  In my case I think FCP was complaining about the ':' character.  That folder is hidden and used by Dockers that are Linux based.  That folder shouldn't even be browsed or manipulated by Windows.  Can that one be skipped for that check?

There's just no satisfying you is there  ;)  Curious though what container its from...

 

I can be a PITA at times.  I'm a stickler for detail.  Does it show?

 

MineOS:

/mnt/user/appdata/minecraftos/archive/hexxit/server-hexxit_2016-06-12_14:20:46.tar.gz
/mnt/user/appdata/minecraftos/backup/hexxit/rdiff-backup-data/access_control_lists.2016-06-12T14:20:29Z.snapshot
/mnt/user/appdata/minecraftos/backup/hexxit/rdiff-backup-data/current_mirror.2016-06-12T14:20:29Z.data
/mnt/user/appdata/minecraftos/backup/hexxit/rdiff-backup-data/error_log.2016-06-12T14:20:29Z.data
/mnt/user/appdata/minecraftos/backup/hexxit/rdiff-backup-data/extended_attributes.2016-06-12T14:20:29Z.snapshot
/mnt/user/appdata/minecraftos/backup/hexxit/rdiff-backup-data/file_statistics.2016-06-12T14:20:29Z.data.gz
/mnt/user/appdata/minecraftos/backup/hexxit/rdiff-backup-data/mirror_metadata.2016-06-12T14:20:29Z.snapshot.gz
/mnt/user/appdata/minecraftos/backup/hexxit/rdiff-backup-data/session_statistics.2016-06-12T14:20:29Z.data

 

 

Link to comment

Suggestion:  You are looking in the appdata folder for characters that Windows doesn't like.  In my case I think FCP was complaining about the ':' character.  That folder is hidden and used by Dockers that are Linux based.  That folder shouldn't even be browsed or manipulated by Windows.  Can that one be skipped for that check?

Done.  Illegal characters in file names are now skipped on appdata/CA Backups/any additional excluded folders.  Still performed everywhere else

 

Is there a way to clear the previous results so I can start a new test fresh?

Done.  Old results deleted prior to beginning a new scan  (btw, if all those dupes were within the a CA Backup folder, then you've done backups to more than one disk.  Switch the destination to user folders since you're on 6.2

Would it be possible for this check to provide info on which disks have the files? It's not too hard to find on your own, but since the plugin already knows......

Done.. The disks the file(s) are on are itemized now

Today when my weekly extended tests ran, I discovered that the plugin really doesn't like AD permissions... :) 

Not done.  Thought about it (and ran the extended test a couple times) and decided that this is going to need an HTML readmore or something on the results, as limiting the results to an arbitrary number isn't fair to other users (and unfortunately AD is something of an edge case here).  And I don't want to force those users to fix the first 10 errors, and then have to re-run the tests (takes my server 2 hours to run through the nearly 5.5M files) to see if there are any more....

 

I'll get around to it, but not on this round of update

 

 

 

Link to comment

Suggestion:  You are looking in the appdata folder for characters that Windows doesn't like.  In my case I think FCP was complaining about the ':' character.  That folder is hidden and used by Dockers that are Linux based.  That folder shouldn't even be browsed or manipulated by Windows.  Can that one be skipped for that check?

Done.  Illegal characters in file names are now skipped on appdata/CA Backups/any additional excluded folders.  Still performed everywhere else

 

Is there a way to clear the previous results so I can start a new test fresh?

Done.  Old results deleted prior to beginning a new scan  (btw, if all those dupes were within the a CA Backup folder, then you've done backups to more than one disk.  Switch the destination to user folders since you're on 6.2

Would it be possible for this check to provide info on which disks have the files? It's not too hard to find on your own, but since the plugin already knows......

Done.. The disks the file(s) are on are itemized now

Today when my weekly extended tests ran, I discovered that the plugin really doesn't like AD permissions... :) 

Not done.  Thought about it (and ran the extended test a couple times) and decided that this is going to need an HTML readmore or something on the results, as limiting the results to an arbitrary number isn't fair to other users (and unfortunately AD is something of an edge case here).  And I don't want to force those users to fix the first 10 errors, and then have to re-run the tests (takes my server 2 hours to run through the nearly 1M files) to see if there are any more....

 

I'll get around to it, but not on this round of update

 

Nice job!  I have a feeling the duplicate files are left over from my conversion of disk at a time from ReiserFS to XFS.  I don't think I did it properly back then.  My data has also progressed from v4.x to 6.2, so something along the way might have gotten botched.  Thank you for a tremendous utility to check this stuff.

 

I would not do anything with AD.  I don't think it is used enough to warrant the investment of your time.

Link to comment
  • 2 weeks later...

Good morning.  I am getting very slow write speeds after upgrading to 6.2.  I have tried to install this plugin but after installation  I click on the plugin, it sends me to the settings page.  Unfortunatly the page is blank.  Clicking on settings again show me the normal setting page.  No icon for the new plugin.  Any help would be appreciated.

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.