unRAID Server Release 5.0-rc8a Available

[WeeboTech]

I'm almost in the thought of.. If limetech wants to release 5.0 final with a plugin manager, so be it.

Those who want early adoption just run RC8a.

 

Although speaking frankly, Release candidate, means almost ready for release without any major changes.

So Plugin manager update via the /boot/extra directory would be fine.

 

[prostuff1]

As Joe said, I would prefer to see the Plugin manager as it's own plugin... or something like it.  That way we the community can change and update it outside of the regular unRAID releases.

 

Please Tom, reach out to some of the members here that are active in developing the plugins and have worked on the development of unMenu packages and features.  Bubba, Joe L. WeeboTech, Rajahal, Influencer, speeding_ant, myself, and i am sure I am missing some could offer some input on this.

[JackBauer]

I used a 4.7RC earlier, and in the end it was what was released.

 

I'm getting that level of confidence to move to 5.0 RC8a.

 

I'm pretty happy about that. 

 

(Oh, in plug in manager can be in 5.1, or 5.01, or whatever, but probably is good to just get something out there that is official, stable, etc...)

[madburg]

I have mentioned this before in the announcement's posts. 5.0 needs to have a plugin manager, that's how we run add-ons in a supported manner.

 

I will explain why this is so. Quoting email so nothing is lost.

 

Email communication with Tom back on 7-12-2011:

 

Me: "I would like to throughout a request, most of the NAS appliances/software out there that I know of (ReadyNAS, FreeNAS as examples) come with an email client and one only has to enter details in a configuration page to send out alerts and would be great if you could add one as well. This way is would be certified by you, so everyone out there would not have to muck with different plugins to get emails from unRAID. Because out there right now there are “add C compiler, makepackage, etc. ssmtp, openssl, mailx”. I am not saying they may or may not be needed but it’s a black hole for many of us, we don’t know if it’s the proper version, etc… Having mail (knowing remotely there is an issue or condition and an automatic action can be taken on ones behalf, like overheat condition) and UPS support (gracefully shutdown in case of a power lose) are primary and lower everyone chances of failures and support . All others are secondary. I am sure the community can build the scripts but the core installed modules should come from you."

 

Tom: "As for email – yes, this has been brought up many times and should have been included early on – just something I haven’t quite been able to get to yet.  Instead I’m taking an approach of designing an “official” plugin architecture where you will be able to easily download and install additional functionality, produced both by Lime Tech as well as other experienced authors.  In fact this is the main point of 5.0 – just trying to get to that point."

 

Now to be fair, he has the right to change his mind. But the right thing to do is to share that with us, either way.

 

I bought into version 5 (knowing it was beta), I had no idea it would take this long to get to a final (which is still an unknown). Those who bought into 4.x have a good return on their investment. Right now its very hard to try to put together a large recommend system for 4.x as the recommend parts don's exist or very hard to find, etc... and it does not support newer hardware.

 

I have no problem if unRAID did not support everything, what I mean is say realtek is a pain, drop it then, LSI a pain drop it, AFP/NFS (even though AFP was slated for 5.0), whatever, pick it up in the next version. There are many products that don't support everything. Yet I have not seen any updates in regards to what was slated to be in 5.0. Big plus for GPT addition, but that had to come either way in order to support larger drives (purpose for a NAS). With that, newer controller support (not like LSI controllers are new, or we are running the latest and greatest), they go hand in hand. So where is the email, UPS, plugin manager, web gui (that actually works)? What happened to the AFP addition that would let us store the apple db, etc.. stuff on one drive or area...? Ex. we wait so long for RC6 to RC8 just for the linux community to add the mpt2sas fix, why not give us these other additions in the mean time to start test? How is the WebGUi been getting any better through these releases? I don't see the improvements. Clearing disks, new permissions, mover, none of them respond back via the Web GUI, I consider this core functionality. I tried to delete an unRAID  user account in RC8 one deleted fine the other refused, no explanation as to why, until I stopped the array, then the second account deleted (well what about all the permissions set on data via that account?). This is core functionality, it should work. if emhttp is not working out for core functionality, I say hell with it, dump it and give us the replacement to start testing with. Leave emhttp with 4.x

 

Forgive me here, but if I buy a new piece of new hardware, I don't need to load a new version of windows. I just check there is a certified driver available and install it. At this pace, some new hardware will come up and need to be added, tested, etc... there will never be a final. Something needs to change to have unRAID version progress faster. I just don't understand why all the core stuff has not been completed or at the very least introduced for testing.

 

Big gripe, many people post bugs/requests in the "unRAID OS 5.0-rc" thread as requested by Tom, but I don't see Tom replying or acknowledging any in a long time. If you start something stick to it. Mods can help a bit if they believe its an error on the part of the user, but Tom should be stepping in and acknowledging if something is a bug or not.

 

There are road map posts well over a year old, never updated, deleted them. Bug tracking posts all over the place, none maintained, delete them all.

 

 

[WeeboTech]

Before this gets into a long discussion, can we keep this thread specific to this release candidate and problems with it?

The tangent discussions are all valid topics, but should probably be separate threads so this one stays concise with what Tom needs in order to move forward with this release candidate.

Post a new thread, possibly drop in a pointer here to it.

 

[SCSI]

Finally finished the update from b12a to RC8a and so far so good.

 

Now I get SG_IO: bad/missing sense data, sb.. on my console and Unraid always think my cache drive is spun-down (blinking green) on the device status page. My cache drive is working fine though. I don't get the error messages if I revert to 12a.

 

My setup is unRAID on ESXi with my cache drive as RDM (Raw Device Mapping)

[Ostrich]

I am wanting to upgrade to this latest RC, as I am about to start purchasing 3TB drives and need the support (currently on 4.7).

 

Just curious about what the following does (taken from upgrade instructions), and impact is if i dont do it?

"Go to Utils/New Permissions and execute that utility to change file ownership and permission settings. This is necessary for proper operation of the 5.0 security model. "

 

 

[Influencer]

It changes the permissions of all of your files to nobody:users from root:root.

 

Impact if you don't do it, you most likely won't be able to access any of your shares over the network unless you only use the root user.

 

^^ I know it will cause issues with accessing your shares, accessing with root user *may* work, and *may* not, I haven't read of anyone doing so and haven't tried that myself.

[jumperalex]

For reasons unrelated to actually testing this release, and not because I have an array problem, I attempted to SSH into the server while the array was in maintenance mode.  But I could not connect using FireSSH.  With the array mounted normally it works fine.  I thought the whole point was to be able to issue commands like reiserfsck (see below) ... can I not SSH in but only telenet?  Do I need to work from the server itself vice a remote session?  Or did I find a bug?

 

There is also a new feature which is a checkbox under the Start button on the Main paged labeled "Maintenance mode".  If checked, when you click Start, the server will come up in "Maintenance Mode".

 

What Maintenance Mode does is bring up the unRaid driver and create the 'md' devices, but does not mount any disk drives (and hence they are also not exported via SMB/NFS/AFP).

 

The main use of this is to let your manually run 'reiserfsck' on your disk drives.  For example, after starting in Maintenance Mode, you can open a telnet session and check the file system of "disk1" using this command:

[JustinAiken]

I don't think SSH has ever worked out of the box...

[jumperalex]

I don't think SSH has ever worked out of the box...

 

Yes but it does work when you install the plugin like I did and works while the array is online.

 

Sent from my Galaxy Nexus using Tapatalk 2

[Joe L.]

I don't think SSH has ever worked out of the box...

 

Yes but it does work when you install the plugin like I did and works while the array is online.

 

Sent from my Galaxy Nexus using Tapatalk 2

Then (apparently) we need to not use the normal plugin installation for the SSH package (or any other needed in maintenance mode), as it looks like you discovered the normal plugin installation of ADD-ONS is NOT performed when in maintenance mode.  (This probably done to eliminate issues that would occur if a package was causing problems, or if a add-on depended upon mounted drives)

 

Send an e-mail to the plugin creator and ask guidance.

 

Joe L.

[jumperalex]

I don't think SSH has ever worked out of the box...

 

Yes but it does work when you install the plugin like I did and works while the array is online.

 

Sent from my Galaxy Nexus using Tapatalk 2

Then (apparently) we need to not use the normal plugin installation for the SSH package (or any other needed in maintenance mode), as it looks like you discovered the normal plugin installation of ADD-ONS is NOT performed when in maintenance mode.  (This probably done to eliminate issues that would occur if a package was causing problems, or if a add-on depended upon mounted drives)

 

Send an e-mail to the plugin creator and ask guidance.

 

Joe L.

 

hmmm I will do that (though I bet he sees it here ) but I think your assertion about WHY is a good one and we might NOT want the plug-in loaded. I mean if you're in maintenance mode then either telnet (if not disabled like on my system) or local console are the correct solutions vice depending on a plug-in that might be part of your problem in the first place. 

 

But even that discussion belongs in another thread, so I'll go do that   Thanks

[dgaschk]

It changes the permissions of all of your files to nobody:users from root:root.

 

Impact if you don't do it, you most likely won't be able to access any of your shares over the network unless you only use the root user.

 

^^ I know it will cause issues with accessing your shares, accessing with root user *may* work, and *may* not, I haven't read of anyone doing so and haven't tried that myself.

 

The root user is prohibited from accessing file shares in the latest RCs.

[dalben]

I'm not sure if the right people have noticed it but there is a very real problem with RC8 detailed here: http://lime-technology.com/forum/index.php?topic=22647.0

[StevenD]

Now I get SG_IO: bad/missing sense data, sb.. on my console

 

I'm getting the same error on my console.  Nothing shows up in the log.  I did not have this on RC5.

[optiman]

Now I get SG_IO: bad/missing sense data, sb.. on my console

 

I'm getting the same error on my console.  Nothing shows up in the log.  I did not have this on RC5.

 

I just checked and I do not have this error message on my console and I'm running RC8.  Could it be specific to your hardware?

[optiman]

I'm not sure if the right people have noticed it but there is a very real problem with RC8 detailed here: http://lime-technology.com/forum/index.php?topic=22647.0

 

wow, that looks to be serious.  That said, I don't seem to have that problem.  I don't have any tmp files to delete, and I'm running win7 64bit as my main client, connecting via SMB shares.  I have deleted files and they remain gone, and no tmp files.  Could it be a plugin issue and not directly related to RC8?

 

What is the easist way to confirm if you are impacted by this issue, just incase  I'm missing something?

 

[dalben]

What is the easist way to confirm if you are impacted by this issue, just incase  I'm missing something?

 

I realised it pretty quickly as I have a 3rd party tool that syncs a large number of my directories to a backup NAS.  The sync tool couldn't handle these tmp files so it threw up errors which I looked into.

 

I disabled my sync tool as a first step with no luck.

 

Could it be a plugin ? Maybe, but the problem goes away when I regress back the previous publicly available beta.  That would indicate that there is an issue related to RC8.  Even if it is a plugin or something not core causing the issues, RC8 is triggering it, something the previous release candidates and betas didn't

[dgaschk]

Plugins may not be compatible with newer versions. It's up to the plugin author to update their product for compatibility. Do not expect an unRAID update to address a plugin compatibility issue.

[dalben]

Plugins may not be compatible with newer versions. It's up to the plugin author to update their product for compatibility. Do not expect an unRAID update to address a plugin compatibility issue.

 

If you're replying to me, I don't believe the issue I and others are experiencing is plugin related. It seems t be the version Smb rc8 is using.

[dgaschk]

Plugins may not be compatible with newer versions. It's up to the plugin author to update their product for compatibility. Do not expect an unRAID update to address a plugin compatibility issue.

 

If you're replying to me, I don't believe the issue I and others are experiencing is plugin related. It seems t be the version Smb rc8 is using.

 

This is easy to verify. Does the problem exist on a clean install? See my sig to revert to a clean install.

[Joe L.]

Plugins may not be compatible with newer versions. It's up to the plugin author to update their product for compatibility. Do not expect an unRAID update to address a plugin compatibility issue.

There is an outstanding bug in most versions of unRAId all the way back to the first version with user-shares in dealing with newly created temp files on a drive being deleted  properly when renamed to have the same name as a file on another physical disk.  This was supposedly fixed in 5.0-rc6-r8168-test

 

Perhaps we are seeing that either the same fix is not in rc8a, or, it introduced a different bug.

 

Changes from 5.0-rc5-r8168 to 5.0-rc6-r8168-test

------------------------------------------------

- shfs: fix file rename bug where 'oldpath' needs to be unlinked if 'newpath' is on different disk

 

I doubt this is caused by a plugin, but instead exposed (made evident as a bug) by any number of programs that create temp files and then rename them once the copy is complete.  "rsync"works that way, as does teracopy, and many others.

 

Joe L.

[dalben]

There is an outstanding bug in most versions of unRAId all the way back to the first version with user-shares in dealing with newly created temp files on a drive being deleted  properly when renamed to have the same name as a file on another physical disk.  This was supposedly fixed in 5.0-rc6-r8168-test

 

Perhaps we are seeing that either the same fix is not in rc8a, or, it introduced a different bug.

 

Changes from 5.0-rc5-r8168 to 5.0-rc6-r8168-test

------------------------------------------------

- shfs: fix file rename bug where 'oldpath' needs to be unlinked if 'newpath' is on different disk

 

I doubt this is caused by a plugin, but instead exposed (made evident as a bug) by any number of programs that create temp files and then rename them once the copy is complete.  "rsync"works that way, as does teracopy, and many others.

 

Joe L.

 

Is there a thread about the bug you're talking about Joe ? I've seen that issue on my server caused by Picasa.  A long list of duplicate file errors and I don't know how to resolve the problem.

[jaybee]

As far as I can see, only one person has reported this SAMBA share .tmp file/deletion problem? This is core functionality, and if it were a problem surely more people would have spotted this and reported it? Dalben, can you perform the above tests to confirm it is not a plugin issue first, and then if the latest SAMBA does fit it?

Why does nobody else see this problem? Using NFS instead?