unRAID Server Release 5.0-beta10 Available

[Chris Pollard]

did something change with the mover script in this release?  It seems to be taking forever to move a load of the .AppleDouble type files from my cache drive and I don't ever remember it doing this before.  Its moved around 10gb of files but its taken hours.

 

afp also seems to have gone a little wonky since beta8d,  my laptop no longer seems to be able to connect whereas my minis both seem to be fine.

 

Lots or errors in the log relating to afp too.

 

syslog.txt

[prostuff1]

did something change with the mover script in this release?  It seems to be taking forever to move a load of the .AppleDouble type files from my cache drive and I don't ever remember it doing this before.  Its moved around 10gb of files but its taken hours.

 

afp also seems to have gone a little wonky since beta8d,  my laptop no longer seems to be able to connect whereas my minis both seem to be fine.

 

Lots or errors in the log relating to afp too.

 

 

Yes, mover did change and it no longer ignores folders with a leading "." (period).  Read threw the thread and you should be able to find it.

[Chris Pollard]

did something change with the mover script in this release? 

 

Yes, mover did change and it no longer ignores folders with a leading "." (period).  Read threw the thread and you should be able to find it.

 

OK I'll wait for that to sort itself out then,  could be thats what is causing my laptop grief.  Thanks.

 

It would be nice to see all the changes in the change log, might save some noise in the release threads.

[abeta]

I'm currently on unRaid 4.7. It's been working fine, but since I recently made a backup of all of my data, I figured I'd try the Beta 10.

 

I followed the instructions. Copied the bzroot and bzimage file over, deleted the config/passwd and config/smbpasswd files. I didn't have a config/shadow file to delete.

 

How long should it take to come up? With 4.7 it takes a minute or two (not long at all). I waited for 5.0Beta10 to come up and on the webpage, it looks like part of the new unRaid screen, but nothing is visible. It is like it is missing the graphics/images, etc.

 

I'm on Win7 64bit and IE9. I've tried Safari on the Mac, Firefox on both PC and Mac and it doesn't ever seem to come up. This is after waiting upwards of 30 minutes. I've copied the files over again, etc. Same thing.

 

I put the old config files and 4.7 system files and it comes up immediately for me. Am I missing something really basic in the upgrade instructions?

 

This is on Lime Tech's MD-1510 server.

 

Sounds like you have "php" loaded as an add-on.   

 

Aye, I don't recall why I installed it . Any idea if  unmenu, airvideo, apcupsd or powerdown require it? Those are probably the only packages I can't live without at this point. If it does, I'll just sit on the sidelines and wait .

You can fix it in about 30 seconds.  

 

See here:

http://lime-technology.com/forum/index.php?topic=10840.msg103297#msg103297

 

however, none of the packages you mentioned use "php"

 

I put it in the go script and still no go. The unmenu stuff seems ok I can see everything fine but I can't start the array .

 

The unRaid screen is still garbage like before. Just a little mountaintop with blank buttons on the top left and right. I can telnet to the box just fine.

 

[prostuff1]

Sounds like your browser is doing some web page caching.  Hold shift and hit the refresh button in your browser; do that a couple times and the issue should clear.

[abeta]

Sounds like your browser is doing some web page caching.  Hold shift and hit the refresh button in your browser; do that a couple times and the issue should clear.

 

No luck with Firefox, Chrome, IE9, Safari on 2 different PCs. Firefox, Chrome and Safari on a MAC. Same problem.

[dgaschk]

Please post a screen shot.

[speeding_ant]

Is it normal for drives to spin up when you access the webgui? Has done it on three different hardware configurations since 5 beta6.

 

Cheers

[SSD]

Is it normal for drives to spin up when you access the webgui? Has done it on three different hardware configurations since 5 beta6.

 

Cheers

 

No - not normal.  Please post these types of questions in the general support forum (Thanks!)

[limetech]

The mover schedule is placed in /etc/cron.d/root ... from the logfile:

Aug 1 00:39:56 Tower login[9423]: ROOT LOGIN on '/dev/pts/0' from '10.2.1.15'
Aug 1 00:41:29 Tower emhttp: shcmd (77): crontab -c /etc/cron.d - <<< "# Generated mover schedule: 0 */2 * * * /usr/local/sbin/mover |& logger"

 

I will investigate the contents of /etc/cron.d next time my server is rebooted.

 

I can confirm that, after a reboot, /etc/cron.d is empty.

 

If 'Apply' is clicked in Mover Settings, then the file /etc/cron.d/root is created with the mover schedule in it.

 

This, clearly, has to be a bug.

 

Yes, fixed in next beta.

[limetech]

did something change with the mover script in this release?  It seems to be taking forever to move a load of the .AppleDouble type files from my cache drive and I don't ever remember it doing this before.  Its moved around 10gb of files but its taken hours.

 

afp also seems to have gone a little wonky since beta8d,  my laptop no longer seems to be able to connect whereas my minis both seem to be fine.

 

Lots or errors in the log relating to afp too.

 

 

Yes, mover did change and it no longer ignores folders with a leading "." (period).  Read threw the thread and you should be able to find it.

 

Mover changed somewhat, but still ignores top-level directories on Cache drive which start with a '.' character.

[limetech]

Status Update

 

I've spent most of last weekend and last couple of days trying to chase down problems integrating the latest linux kernel 3.0.0.  This is in an effort to fix some nagging problems with slow Parity Sync/Check, shutdown kernel crashes, and a few other nits.  But this has been very frustrating....

 

Here's a brief history of unRaid releases vs. linux kernel releases:

 

5.0-beta9 / 2.6.37.6  <= last "solid" kernel with unRaid

5.0-beta10 / 2.6.39.3  <= above mentioned issues

5.0-beta11 / 3.0.0  <= similar issues as above plus a few others

(-beta11 has not been released)

 

(I skipped 2.6.38.x because of other problems I can't remember at the moment.)

 

One might ask: If 2.6.37.6 was so solid, why upgrade?  Well I always monitor kernel change logs, looking for changes in storage and network drivers.  When I see a significant bug fixed or improvement, then it's time to upgrade.  But between 2.6.37.6 and 3.0.0 something fairly radical changed in the kernel which has to do with I/O scheduling, and I haven't been able to isolate this yet after many hours of testing and debugging over the last several days.

 

So...... Here's the plan: I'm going to release -beta11 reverting to kernel 2.6.37.6 for now, with latest netatalk and some Active Directory fixes and a plugin manager working so I can move on and get 5.0-rc1 out.  This will let 3.0.0 bake for a while longer and when I turn attention back to upgrading the kernel, 3.1.0 (or 3.0.1 ?) will probably be out.

[speeding_ant]

Nice one - cheers for the updates as per usual

[Interstellar]

Cool - look forward to it.

 

 

Also, I think speeding_ant, the S3 you suggested to me doesn't work properly anymore 

 

It goes to sleep when my Mac Pro is awake (192.168.1.5) and it's supposed to be pinging that.

 

I'll post my script in my original thread on S3 sleep!

[PeterB]

5.0-beta9 / 2.6.37.6   <= last "solid" kernel with unRaid

5.0-beta10 / 2.6.39.3  <= above mentioned issues

5.0-beta11 / 3.0.0  <= similar issues as above plus a few others

(-beta11 has not been released)

 

(I skipped 2.6.38.x because of other problems I can't remember at the moment.)

 

Wasn't it 2.6.38 which first included the '100 line patch' to improve desktop responsiveness?  Is it possible that this is what is screwing with unRAID I/O?

[pfp]

Just added 24 drive support in unreleased 4.5.1 and working with a guy to get SAS controllers working.

 

Whatever happened to 24 drive support?

[abeta]

Please post a screen shot.

 

Attached.

[speeding_ant]

Whatever patch you applied to allow PHP short tags didn't work. Post the tag you put in the go script.

 

Also, have you restarted your server since you added it to the go script?

[dimaestro]

And a dead server again today. Gonna drop back to 4.7. I just hope the file copies I initiated over lunch completed correctly.

[abeta]

Whatever patch you applied to allow PHP short tags didn't work. Post the tag you put in the go script.

 

Also, have you restarted your server since you added it to the go script?

 

I cut/pasted from the thread indicated

 

sed -i "s/^short_open_tag = Off/short_open_tag = On/" /boot/custom/php/php.ini

 

When I cut/paste into the terminal window:

 

login as: root

root@Serenity's password:

Linux 2.6.32.9-unRAID.

root@Serenity:~# sed -i "s/^short_open_tag = Off/short_open_tag = On/" /boot/custom/php/php.ini

sed: can't read /boot/custom/php/php.ini: No such file or directory

 

There's no php directory in custom.

 

 

root@Serenity:/boot# cd custom

root@Serenity:/boot/custom# ls

etc/  transmission/

root@Serenity:/boot/custom#

 

I don't believe Transmission is actually running at this point.

 

[ftp222]

Just added 24 drive support in unreleased 4.5.1 and working with a guy to get SAS controllers working.

 

Whatever happened to 24 drive support?

 

I was curious about the same thing - any chance this gets added into the 5.0 release?

 

Please

 

[rlung]

When using user share & fill-up method whenever any drive dips below threshold, windows will report both incorrect total size & free space on mapped share

 

16 drives

fill up

25,000,000 min free space

split 0

 

when any drive is > 25 gb windows reports 12.9 TB / 1.87 TB ... total / used space (the correct amount)

any drive dips < 25; > 24 windows reports 12.0 / 1.78

any drive dips < 24; > 23 windows reports 10.6 / 1.74

any drive dips < 23; > 22 windows reports 9.77 / 1.69

any drive dips < 22; > 21 windows reports 2.72 / 1.57 (2.72 not a typo)

 

Once drive space increases to above 'min free space' threshold normal readings appear.

 

Incorrect reading continued on another unraid 5.0b10 server using split 3.

 

Bother servers migrated from a 4.7 environment with no change in settings and not displaying this behavior.

 

- correction there is 1 change; both servers currently have no parity drive set. They did in the 4.7 environment. Will set parity, rebuild and observe

[toby9999]

When using user share & fill-up method whenever any drive dips below threshold, windows will report both incorrect total size & free space on mapped share

 

16 drives

fill up

25,000,000 min free space

split 0

 

when any drive is > 25 gb windows reports 12.9 TB / 1.87 TB ... total / used space (the correct amount)

any drive dips < 25; > 24 windows reports 12.0 / 1.78

any drive dips < 24; > 23 windows reports 10.6 / 1.74

any drive dips < 23; > 22 windows reports 9.77 / 1.69

any drive dips < 22; > 21 windows reports 2.72 / 1.57 (2.72 not a typo)

 

Once drive space increases to above 'min free space' threshold normal readings appear.

 

Incorrect reading continued on another unraid 5.0b10 server using split 3.

 

Bother servers migrated from a 4.7 environment with no change in settings and not displaying this behavior.

 

- correction there is 1 change; both servers currently have no parity drive set. They did in the 4.7 environment. Will set parity, rebuild and observe

This is very interesting ... I'm just starting to play with this feature with large numbers of HDDs ... I'll need to keep a close eye on this and see if I get the same problem ... thanks for the warning (and hopefully Tom can find the bug if there is one).

[purko]

...

5.0-beta9 / 2.6.37.6   <= last "solid" kernel with unRaid

5.0-beta10 / 2.6.39.3  <= above mentioned issues

5.0-beta11 / 3.0.0  <= similar issues as above plus a few others

...

 

Is there any chance that the stable 4.7.1 could get that "solid" 2.6.37.6 kernel?

(I'm only allowed to put stable releases on our production servers)

 

 

[rlung]

When using user share & fill-up method whenever any drive dips below threshold, windows will report both incorrect total size & free space on mapped share

 

16 drives

fill up

25,000,000 min free space

split 0

 

when any drive is > 25 gb windows reports 12.9 TB / 1.87 TB ... total / used space (the correct amount)

any drive dips < 25; > 24 windows reports 12.0 / 1.78

any drive dips < 24; > 23 windows reports 10.6 / 1.74

any drive dips < 23; > 22 windows reports 9.77 / 1.69

any drive dips < 22; > 21 windows reports 2.72 / 1.57 (2.72 not a typo)

 

Once drive space increases to above 'min free space' threshold normal readings appear.

 

Incorrect reading continued on another unraid 5.0b10 server using split 3.

 

Bother servers migrated from a 4.7 environment with no change in settings and not displaying this behavior.

 

- correction there is 1 change; both servers currently have no parity drive set. They did in the 4.7 environment. Will set parity, rebuild and observe

This is very interesting ... I'm just starting to play with this feature with large numbers of HDDs ... I'll need to keep a close eye on this and see if I get the same problem ... thanks for the warning (and hopefully Tom can find the bug if there is one).

 

Parity rebuild completed, behavior persists.

 

Narrowed and can duplicate on demand.

 

This test is on another server running 5b10:

 

20 data + 1 parity, no cache

1 and only 1 user share labeled 'ALL' utilizing all disks available

allocation = fill up

min free space = 30,000,000

split = 3

 

free space on all 20 disks are > 30gb; windows correctly shows 32.7 tb/ 4.58 tb (total / free)

 

1st disk written with 4.35gb file to < 30 gb; windows reports 31.3 / 4.55 (disk 1 = 1.5 tb)

- any further writing to this disk does not affect either value beyond the expected

2nd disk written with 4.35gb to < 30 gb; windows reports 30.0 / 4.52 (disk 2 = 1.5 tb)

- further writing to disk no abnormal change

3rd disk written with 4.35gb to < 30gb; windows reports 28.7 / 4.49 (disk 3 = 1.5 tb)

- further writing no abnormal change

 

Data concludes that user shares is discarding total / free space of individual disks as they fall below min free space threshold. Able to duplicate this without writting by artificially raising / lower 'min free space' of user share and noting what windows displays.

 

update 08/15/11

 

5.0beta11 fixed above