unRAID Server Release 5.0-beta7 Available

[lionelhutz]

Tom - Nice work on the 3T support. I'm waiting to hear how bjp999 gets along with the 3T drives.

 

 

 

I upgraded to 5.0 beta 7 and everything was working fine.  I then added some unmenu plugins and now I can't access the unraid main page, its all blank.

 

Please post issues caused by add-ons in the user customization forum. This thread is for issues related to the unRAID release itself.

 

Peter

 

[Thomss]

Does anybody know why a share which I made is no accessible in windows after updating from 4.7 to 5.07?

 

 

"You do not have permission to access \\unraid\backups. contact your network admin to request access"

 

I can see in the unraid gui the files are there and the settings are the same as other shares which I have and can see in windows.

 

i have done the permissions thing in the unraid gui.

[BRiT]

Tom, which Slackware version this build is this version based on?

 

The closest is Slackware 13.1, where as the Linux kernel matches Slackware 13.37.

[gfjardim]

I guess still based on Slack 13.1 because of gcc version 4.4.4, while Slack 13.37 has gcc version 4.5.2.

 

 

[SSD]

Continued ...

 

Parity build completed successfully. 

 

I stopped the array, removed the 2.2T disk from the array, then started the array.

 

Disk was simulated as expected.

 

Stopped the array, assigned a 3T precleared disk to the simulated slot, and started the array.

 

emhttp got to "mounting" stage and then froze for several minutes.  I thought it was hung.  But finally it did come back (20G of the disk had been rebuilt when emhttp finally returned to give some idea of how long it took).

 

Disk continues to be simulared correctly.  Rebuild is running.

 

All working perfectly so far.    Will post back when rebuild completes.

 

Here is the GPT partition created on the precleared disk (remember preclear tries to put a MBR and partition on the disk, but since the disk is too big for an MBR, it can't really do that properly.  That didn't create any problem for unRAID, which replaced the MBR with the GPT):

 

root@Tower:~# gdisk /dev/sdd

GPT fdisk (gdisk) version 0.6.14

 

Partition table scan:

  MBR: protective

  BSD: not present

  APM: not present

  GPT: present

 

Found valid GPT with protective MBR; using GPT.

 

Command (? for help): i

Using 1

Partition GUID code: EBD0A0A2-B9E5-4433-87C0-68B6B72699C7 (Linux/Windows data)

Partition unique GUID: AC460187-C95B-4D5F-9ED8-433DDFA91A1C

First sector: 64 (at 32.0 KiB)

Last sector: 5860533134 (at 2.7 TiB)

Partition size: 5860533071 sectors (2.7 TiB)

Attribute flags: 0000000000000000

Partition name:

[PeterB]

At long last, my new WD20EARS has arrived.  I presume that there is no reason why I shouldn't use GPT on it in order to start testing the new functionality?

[SSD]

At long last, my new WD20EARS has arrived.  I presume that there is no reason why I shouldn't use GPT on it in order to start testing the new functionality?

 

I belieive the GPT partitions are only created for drives >2TiB.  You can't control which drives are GPT and which are MBR.

[Joe L.]

At long last, my new WD20EARS has arrived.  I presume that there is no reason why I shouldn't use GPT on it in order to start testing the new functionality?

I don't think you get to choose.  

All I see with 2TB drives as choices are MBR-4k-aligned, or MBR-un-unaligned.

[jaybee]

Can someone point me to where I can read what each 5.0 beta release contains without having to go into the opening post of each beta thread and read everything?

 

If you read the release notes posted in the first message of this thread, you will see all the significant changes between 4.5.6 and the current 5.0beta7.

 

Unfortunately, because 5 was forked off before the current 4.x release, you cannot see the differences between 4.7 and the current 5.0 beta.  However, if you look at the 4.7 release notes, you should find the changes from 4.5.6 and 4.7.

 

Ok thanks for this. Right, so can someone summarize for me why one would want to go for a 5 beta over the 4.7 stable? Is there much risk? I see a LOT of bug fixes and tweaks between 4.7 final and up to now throughout the 5 betas. I want to do a new build but not sure which one to select for my build. I do care that my data stays as stable as possible yes. What are the main juicy reasons if you like to consider a 5 beta over 4.7 final. I mean, list the main NEW features or must have ammendments and/or bug fixes. This would be very helpful for someone like me just about to do their first build.

[PeterB]

At long last, my new WD20EARS has arrived.  I presume that there is no reason why I shouldn't use GPT on it in order to start testing the new functionality?

 

I belieive the GPT partitions are only created for drives >2TiB.  You can't control which drives are GPT and which are MBR.

 

Ah, I thought that

If gdisk  finds  a valid  MBR or BSD disklabel but no GPT data, it will attempt to convert the MBR or disklabel into GPT form.

from man gdisk, meant that any MBR drive could be converted to GPT.

[PeterB]

If you read the release notes posted in the first message of this thread, you will see all the significant changes between 4.5.6 and the current 5.0beta7.

 

I see a LOT of bug fixes and tweaks between 4.7 final and up to now throughout the 5 betas.

 

No, you may see a lot of bug fixes and enhancements between 4.5.6 and the current 5 beta.

 

Several of the bug fixes and a few of the enhancements have also been applied to 4.6/4.7.

 

I want to do a new build but not sure which one to select for my build. I do care that my data stays as stable as possible yes. What are the main juicy reasons if you like to consider a 5 beta over 4.7 final. I mean, list the main NEW features or must have ammendments and/or bug fixes.

 

Well, from a personal point of view, v5 fixes the problem I was experiencing on more than 50% of boots, where the flash drive was not being mounted after bzroot/bzimage loaded, resulting in no network, no array etc.  I have not experience this problem once since moving to 5.0 (and my machine is booted more often than most in this land where power cuts happen, on average, once per day).

 

v5.0beta3 added support for my LSI 2008-based sas/sata interface card.

 

The browser interface on 5.0 is sooo much nicer than the old one.

 

I started running v5.0beta6 st the very beginning of March, and recently upgraded to 5.0beta6d, and on to beta7.  I haven't experienced a single problem with it.

[Thomss]

Any ideas as to why 1 of my shares is not able to be accessed  on a PC/MAC?

[gfjardim]

I'm sure this was not the intention, Beta7 now spinup/spindown the drives in VMWare ESXi using mapped raw disks, but temperatures still a non go. The tests were done using either LSI Logic SAS controller (mptsas driver) and Paravirtual controller (pvscsi driver with recompiled kernel).

 

If you, Tom, could fix this little temperature bug, we can dramatically improve the usability of unRAID, since the users could consolidate router / PVR / home automation / storage appliances into one physical machine.

[sacretagent]

I think i might have found a bug

 

i upgraded 2 days ago ... and i knew i had a disk which was running on it's last feet (a lot of read errors)

so this morning i woke up and the disk was dead ....

so i went to the shop and got another one ...

pulled out the bad one .... put in the new one .... and let it rebuild ....  (no preclear... i know not th best thing to do but i was in a hurry)

 

so now the rebuild has completed but the disk still shows unformatted in the menu...

and if i SSH to /mnt/disk7 then it is empty

 

root@p5bplus:~# cd /mnt/disk7

-bash: cd: /mnt/disk7: No such file or directory

 

so not very sure what is going on now

 

guess i will pull the disk and format it and hope it rebuilds again ?

 

any other ideas?

 

[Joe L.]

I think i might have found a bug

 

i upgraded 2 days ago ... and i knew i had a disk which was running on it's last feet (a lot of read errors)

so this morning i woke up and the disk was dead ....

so i went to the shop and got another one ...

pulled out the bad one .... put in the new one .... and let it rebuild ....  (no preclear... i know not th best thing to do but i was in a hurry)

 

so now the rebuild has completed but the disk still shows unformatted in the menu...

and if i SSH to /mnt/disk7 then it is empty

 

root@p5bplus:~# cd /mnt/disk7

-bash: cd: /mnt/disk7: No such file or directory

 

so not very sure what is going on now

 

guess i will pull the disk and format it and hope it rebuilds again ?

 

any other ideas?

 

un-formatted simply indicates it was not mounted.

 

You might want to run a reiserfsck on the /dev/mdX device

reiserfsck --check /dev/mdX

 

and post a system log... none of us are mind-readers.

 

DO NOT FORMAT THE DISK.  IF YOU DO THE CONTENTS WILL BE GONE!

[sacretagent]

it should mount automatically not after rebuilding??

it didn't do that .... even after reboot ...

put array down....unassigned drive ....

mkreiserfs..... assigned drive to empty slot ...

started array again... and it is rebuilding again.... waiting 8 hours again

 

no syslogs cause of the reboot ...would have posted if i had one ....

 

mmm it doesn't look promising

 

root@p5bplus:/# df -T
Filesystem    Type   1K-blocks      Used Available Use% Mounted on
/dev/sda1     vfat      974304    200300    774004  21% /boot
/dev/sdp1 reiserfs   488371640  12544072 475827568   3% /mnt/cache
/dev/md2  reiserfs   976732736 952704952  24027784  98% /mnt/disk2
/dev/md12 reiserfs   732552188 705455852  27096336  97% /mnt/disk12
/dev/md5  reiserfs   976732736 949951700  26781036  98% /mnt/disk5
/dev/md9  reiserfs   976732736 951217188  25515548  98% /mnt/disk9
/dev/md4  reiserfs   1953454928 1926266528  27188400  99% /mnt/disk4
/dev/md11 reiserfs   732552188 715024024  17528164  98% /mnt/disk11
/dev/md10 reiserfs   976732736 958187864  18544872  99% /mnt/disk10
/dev/md1  reiserfs   976732736 950052400  26680336  98% /mnt/disk1
/dev/md13 reiserfs   976732736 949666996  27065740  98% /mnt/disk13
/dev/md6  reiserfs   976732736 956824180  19908556  98% /mnt/disk6
/dev/md8  reiserfs   976732736 950253392  26479344  98% /mnt/disk8
/dev/md3  reiserfs   976732736 950335124  26397612  98% /mnt/disk3
shfs     fuse.shfs   12209153928 11915940200 293213728  98% /mnt/user

 

no disk 7 to seen anywhere

only /dev/sdb

anyway i can see if that sucker has a filsystem ?

[Joe L.]

it should mount automatically not after rebuilding??

it didn't do that .... even after reboot ...

put array down....unassigned drive ....

mkreiserfs..... assigned drive to empty slot ...

started array again... and it is rebuilding again.... waiting 8 hours again

 

no syslogs cause of the reboot ...would have posted if i had one ....

You did not do as I said. 

I did not want you to stop the array, but to run reiserfsck on the /dev/mdX device with the array running.

 

What you did was to fix the drive outside of the array and basically break parity.  Now, unRAID is re-writing the same exact corrupted file-system it is re-constructing from parity.

 

The process to run reiserfsck on a drive is given in the wiki here:

http://lime-technology.com/wiki/index.php?title=Check_Disk_Filesystems

[sacretagent]

so what you are saying is that i might even loose my data ??

although that the disk rebuilds fine ??

the parity is 14 days old...

it was before the disk started to have issue ..

so a file system errors just gets copied in the parity ?

mmm guessing that is not a good thing ....

remind you that the old disk NEVER got a red ball

and that now after rebuilding everything was green... just the unformatted was there

and no MD7 in df-T or no /mnt/disk7

 

 

 

[limetech]

I think i might have found a bug

 

i upgraded 2 days ago ... and i knew i had a disk which was running on it's last feet (a lot of read errors)

so this morning i woke up and the disk was dead ....

What do you mean "dead"?  Was there a red icon next to it?  Even if the disk is disabled, "disk7" should still show up and be accessible.  While reconstruct is in progress, disk7 should still show up and be accessible.  When the "rebuild" ran, which disk was being written (hopefully disk7).

 

[RobJ]

Can someone point me to where I can read what each 5.0 beta release contains without having to go into the opening post of each beta thread and read everything? I basically want an easy to see list of reasons I should select 4.7 final or the latest stable 5.0 beta.

 

While you have been well answered, and have no need of this, I would like to point to all users the wiki page that contains the complete Release Notes for unRAID.

 

  Release Notes

[Joe L.]

.

remind you that the old disk NEVER got a red ball

Do you still have the disk that did not have a "red" indicator, but was having read errors?  

 

Here is what I'm guessing you did.

A)  You loaded your old disk with data and THEN assigned a parity disk.  The initial parity calc occurred after your disks had been loaded with data. (you did this to be faster)

or

B)  did you assign parity and data, then load the data while automatically calculating parity? (loading data would have been slower with parity calcs occurring as the load occurred)

 

Which one was it? "A", or "B"

[SSD]

Continued ...

 

Rebuild completed successfully.  3T data disk is working well.  Computed md5 on several files from old and new and compared perfectly.

 

As soon as the rebuild completed, I started a parity check.

 

The parity check almost immediately reported 5 sync errors:

 

Jun 10 06:35:05 Shark kernel: md: parity incorrect: 44640 (Errors)

Jun 10 06:35:05 Shark kernel: md: parity incorrect: 44648 (Errors)

Jun 10 06:35:05 Shark kernel: md: parity incorrect: 48680 (Errors)

Jun 10 06:35:05 Shark kernel: md: parity incorrect: 48688 (Errors)

Jun 10 06:35:05 Shark kernel: md: parity incorrect: 48696 (Errors)

 

These parity checks are at the very beginning of the drive, in an area we call the "housekeeping" area.  There have been no other sync errors.  Parity check is about to the 2.6T point, with 400G to go.

 

I can correct these sync errors without much trouble.  Since they are not in the data area they are not a huge worry for me.  Would be more concerning if sync errors creap in during normal use, which I see no evidence so far.  I'll be running frequent parity checks of the housekeeping area to see if any other operations cause this issue.

[jaybee]

Thanks for the suggestions guys. I might go with the 5.X beta stable one, think it was 5.5. I will check. Ta.

[Joe L.]

Thanks for the suggestions guys. I might go with the 5.X beta stable one, think it was 5.5. I will check. Ta.

No 5.X release is stable.   Sorry, they are all beta versions.   5.0beta7 is one of the best so far, but it still has known bugs.

[SSD]

Thanks for the suggestions guys. I might go with the 5.X beta stable one, think it was 5.5. I will check. Ta.

No 5.X release is stable.   Sorry, they are all beta versions.   5.0beta7 is one of the best so far, but it still has known bugs.

 

Stable and bug-free are not the same thing.

 

4.7 is the release version.

 

5.0ba, despite some known bugs, is used by many forum members.  It has proven itself stable (no crashes or data corruptions) once installed successfully.

 

5.0b7 is brand new.  Only a handful of users have tried it.  It has not yet proven itself stable, but appears promising.

 

I would not advise users to load 5.0b7 until there are more thumbs ups from users with test arrays.  But each person can decide what risk they are willing to take, and make the decision of what version to load.