unRAID Server Release 5.0-beta13 Available


limetech

Recommended Posts

  • Replies 269
  • Created
  • Last Reply

Top Posters In This Topic

I have a very basic Supermicro X7SPE-HF-525 with 4 Hitachi 2gb coolspins and everthing is working well.  First time using AFP as well - created a TimeMachine user share, turned on AFP and enabled AFP on the share and disabled SMB and my Lion machine can see it, created the sparsebundle and is backing up now.  So from a "nothing fancy" user things look great.  I chose these components to avoid realtek NICs (hate them) and high compatibility. 

 

Just adding this so people don't get too scared of trying beta 13 (even if you are superstitious)  ;D

Link to comment

Did some testing today.

 

Windows 7 Enterprise as the client.  Set up a 2GB ramdisk on both the Windows client and on unRAID to eliminate disk I/O.  Systems were connected with a crossover cable, and not switch.

 

Both systems are Gigabyte mobos with built-in Realtek NICs.

 

I was able to consistently saturate the wire at 120MB/s.

 

I repeated it with a 5-port Trendnet switch, and got the same results.

 

I added another switch (D-Link, DGS-2208) and got the same results.

 

This is a full 50% improvement over what I could get with exactly the same hardware, using ramdisk-to-ramdisk transfers, under Beta10 (the last one I tested this way).  And that improvement is not just 50%, but giving me full wireline speed.

 

To top it off, COU utilization on the client was about 10%, and on the server it was 50%.

 

Nice job Tom!

 

 

Link to comment

bubbaQ, would it be possible for you to start a new thread going into detail explaining everything that is needed for this? (I ask for a new thread, because I don't want this saturating the 13b thread.)

 

Thanks!

 

Did some testing today.

 

Windows 7 Enterprise as the client.  Set up a 2GB ramdisk on both the Windows client and on unRAID to eliminate disk I/O.  Systems were connected with a crossover cable, and not switch.

 

Both systems are Gigabyte mobos with built-in Realtek NICs.

 

I was able to consistently saturate the wire at 120MB/s.

 

I repeated it with a 5-port Trendnet switch, and got the same results.

 

I added another switch (D-Link, DGS-2208) and got the same results.

 

This is a full 50% improvement over what I could get with exactly the same hardware, using ramdisk-to-ramdisk transfers, under Beta10 (the last one I tested this way).  And that improvement is not just 50%, but giving me full wireline speed.

 

To top it off, COU utilization on the client was about 10%, and on the server it was 50%.

 

Nice job Tom!

 

 

Link to comment

There is not much too it:

 

Create a mountpoint on unRAID for a temporary ramdisk on one of your data disks.

 

mkdir /mnt/disk1/temp

 

Create a ramdisk:

 

mount -t tmpfs tmpfs /mnt/disk1/temp

 

Map a drive from the client to the ramdisk you just created on the server.

 

For a ramdisk on the client, use either:

 

Vsuite Ramdisk Free Edition

http://www.romexsoftware.com/en-us/vsuite-ramdisk/download.html

 

Dataram Ramdisk

http://memory.dataram.com/products-and-services/software/ramdisk

 

Chose an appropriate test file for copying... that will fit on the size ramdisk you created, and that will not exhaust RAM on the unRAID server tmpfs.  Don’t worry about buffers and caching on the server — this test is ONLY to test the wire speed, and not Disk I/O.  However, it should be large enough to take several seconds in order to get an accurate timing... usually 1GB is enough.

 

Copy the file to the ramdisk on the client. 

 

Then copy it from the client’s ramdisk to the server’s ramdisk with whatever tool you want, such as teracopy, or just drag/drop in Windows and use a stopwatch.  Delete it on the client, and copy it back.  Wash, rinse repeat.

 

You can also use the Windows task manager networking tab, and set the update speed to “high.”

Link to comment

 

Does this allow me to combine server 1 that has the Share "Movies" with a second server that also has the share "Movies" to appear as a combined share \\SomeServer\Movies\?

No, but what do you think about that feature (of combining shares from separate servers like that)?

 

Yes, please!! I'd love this to be a possibility!

Link to comment

One thing I noticed with this and the previous beta is that copying files to the directory share with the cache drive enabled is slower than writing directly to the cache drive itself.

 

Test file was 25.4 GB in size and was copied from a solid state drive via Windows 7 64bit. Times given are repeatable. Test file was removed from the cache drive prior to the next test.

 

Share: 6 minutes 10 seconds

Cache: 4 minutes 45 seconds

 

After enabling SMB2 -

 

Share: 5 minutes 57 seconds

Cache: 4 minutes 41 seconds

 

One thing I noticed that when writing to the share, the Windows Task Monitor reported the LAN activity all over the place but was much more stable when writing directly to the cache drive.

 

In the attached pic, the left side is prior to enabling SMB2, the right side is after. Left traffic is directly to the share, right traffic is directly to the cache drive.

Network_Traffic.gif.6330504b2261b7c1260c69c126c67a34.gif

Link to comment

I got all kind of creepy errors with my array, moving back to b12a solved those problems. I've done a reiserfsck on all disks previously the update, so this might be some 3.1.0 kernel weirdness.

 

Attached is the first 5k lines of my syslog, the original file have more than 2 million lines!

 

PS: Johnm, are you using the open-vm-tools package into your VM?

 

OK, today I booted 5b13 outside ESXi and had the same problems before.

 

Oct 31 12:03:19 Servidor emhttp_event: array_started
Oct 31 12:03:19 Servidor kernel: mdcmd (48): check CORRECT
Oct 31 12:03:19 Servidor kernel: md: recovery thread woken up ...
Oct 31 12:03:19 Servidor kernel: md: recovery thread checking parity...
Oct 31 12:03:19 Servidor kernel: md: using 3072k window, over a total of 1953514552 blocks.
Oct 31 12:03:37 Servidor ntpd[1302]: Deferring DNS for a.ntp.br 1
Oct 31 12:03:58 Servidor ntpd[1302]: Deferring DNS for c.ntp.br 1
Oct 31 12:04:19 Servidor sshd[6569]: Accepted password for root from 192.168.0.120 port 49725 ssh2
Oct 31 12:04:19 Servidor sshd[6586]: lastlog_openseek: Couldn't stat /var/log/lastlog: No such file or directory
Oct 31 12:04:19 Servidor sshd[6586]: lastlog_openseek: Couldn't stat /var/log/lastlog: No such file or directory
Oct 31 12:04:20 Servidor ntpd_intres[6527]: host name not found: c.ntp.br
Oct 31 12:04:40 Servidor ntpd_intres[6527]: host name not found: a.ntp.br
Oct 31 12:05:01 Servidor ntpd_intres[6527]: host name not found: c.ntp.br
Oct 31 12:06:00 Servidor root: Running ReiserFS Check on drive 5
Oct 31 12:06:03 Servidor root: Running ReiserFS Check on drive 8
Oct 31 12:06:06 Servidor root: Running ReiserFS Check on drive 10
Oct 31 12:06:23 Servidor ntpd_intres[6527]: host name not found: c.ntp.br
Oct 31 12:06:43 Servidor ntpd_intres[6527]: host name not found: a.ntp.br
Oct 31 12:06:57 Servidor emhttp: Spinning down all drives...
Oct 31 12:06:57 Servidor kernel: mdcmd (49): spindown 0
Oct 31 12:06:57 Servidor kernel: mdcmd (50): spindown 1
Oct 31 12:06:58 Servidor kernel: mdcmd (51): spindown 2
Oct 31 12:06:58 Servidor kernel: mdcmd (52): spindown 3
Oct 31 12:06:59 Servidor kernel: sd 8:0:6:0: [sdj] Device not ready
Oct 31 12:06:59 Servidor kernel: sd 8:0:6:0: [sdj]  Result: hostbyte=0x00 driverbyte=0x08
Oct 31 12:06:59 Servidor kernel: sd 8:0:6:0: [sdj]  Sense Key : 0x2 [current] 
Oct 31 12:06:59 Servidor kernel: sd 8:0:6:0: [sdj]  ASC=0x4 ASCQ=0x2
Oct 31 12:06:59 Servidor kernel: sd 8:0:6:0: [sdj] CDB: cdb[0]=0x28: 28 00 01 a3 7f df 00 00 10 00
Oct 31 12:06:59 Servidor kernel: end_request: I/O error, dev sdj, sector 27492319
Oct 31 12:06:59 Servidor kernel: sd 9:0:2:0: [sdn] Device not ready
Oct 31 12:06:59 Servidor kernel: sd 9:0:2:0: [sdn]  Result: hostbyte=0x00 driverbyte=0x08
Oct 31 12:06:59 Servidor kernel: sd 9:0:2:0: [sdn]  Sense Key : 0x2 [current] 
Oct 31 12:06:59 Servidor kernel: sd 9:0:2:0: [sdn]  ASC=0x4 ASCQ=0x2
Oct 31 12:06:59 Servidor kernel: sd 9:0:2:0: [sdn] CDB: cdb[0]=0x28: 28 00 01 a3 7f e0 00 00 10 00

 

When I tried to run reiserfsck on a sleepy driver, the error was trigged. :(

Link to comment

I upgraded from 4.7 to 5 b13 following all the steps necessary. For some reason I can no longer access that flash drive via the network (error message saying it is unavailable) and one of my shares (Recorded TV) is unavailable and has an orange dot beside it.

I have attached the syslog.

Cheers

 

Reboot the windows PC you are trying to access it from and it will return.

Link to comment

orange dot = there is data for this share on the cache drive.

 

Cheers, always wondered what that was for!

 

Has anyone tried the Supermicro AOC-USASLP-L8I with b13?  I'm running b12, debating whether to upgrade.  Do you need any special kind of switch for the link aggregation? I'm just running an unmanaged netgear gigabit.

Link to comment

I got all kind of creepy errors with my array, moving back to b12a solved those problems. I've done a reiserfsck on all disks previously the update, so this might be some 3.1.0 kernel weirdness.

 

Attached is the first 5k lines of my syslog, the original file have more than 2 million lines!

 

PS: Johnm, are you using the open-vm-tools package into your VM?

 

OK, today I booted 5b13 outside ESXi and had the same problems before.

 

Oct 31 12:03:19 Servidor emhttp_event: array_started
Oct 31 12:03:19 Servidor kernel: mdcmd (48): check CORRECT
Oct 31 12:03:19 Servidor kernel: md: recovery thread woken up ...
Oct 31 12:03:19 Servidor kernel: md: recovery thread checking parity...
Oct 31 12:03:19 Servidor kernel: md: using 3072k window, over a total of 1953514552 blocks.
Oct 31 12:03:37 Servidor ntpd[1302]: Deferring DNS for a.ntp.br 1
Oct 31 12:03:58 Servidor ntpd[1302]: Deferring DNS for c.ntp.br 1
Oct 31 12:04:19 Servidor sshd[6569]: Accepted password for root from 192.168.0.120 port 49725 ssh2
Oct 31 12:04:19 Servidor sshd[6586]: lastlog_openseek: Couldn't stat /var/log/lastlog: No such file or directory
Oct 31 12:04:19 Servidor sshd[6586]: lastlog_openseek: Couldn't stat /var/log/lastlog: No such file or directory
Oct 31 12:04:20 Servidor ntpd_intres[6527]: host name not found: c.ntp.br
Oct 31 12:04:40 Servidor ntpd_intres[6527]: host name not found: a.ntp.br
Oct 31 12:05:01 Servidor ntpd_intres[6527]: host name not found: c.ntp.br
Oct 31 12:06:00 Servidor root: Running ReiserFS Check on drive 5
Oct 31 12:06:03 Servidor root: Running ReiserFS Check on drive 8
Oct 31 12:06:06 Servidor root: Running ReiserFS Check on drive 10
Oct 31 12:06:23 Servidor ntpd_intres[6527]: host name not found: c.ntp.br
Oct 31 12:06:43 Servidor ntpd_intres[6527]: host name not found: a.ntp.br
Oct 31 12:06:57 Servidor emhttp: Spinning down all drives...
Oct 31 12:06:57 Servidor kernel: mdcmd (49): spindown 0
Oct 31 12:06:57 Servidor kernel: mdcmd (50): spindown 1
Oct 31 12:06:58 Servidor kernel: mdcmd (51): spindown 2
Oct 31 12:06:58 Servidor kernel: mdcmd (52): spindown 3
Oct 31 12:06:59 Servidor kernel: sd 8:0:6:0: [sdj] Device not ready
Oct 31 12:06:59 Servidor kernel: sd 8:0:6:0: [sdj]  Result: hostbyte=0x00 driverbyte=0x08
Oct 31 12:06:59 Servidor kernel: sd 8:0:6:0: [sdj]  Sense Key : 0x2 [current] 
Oct 31 12:06:59 Servidor kernel: sd 8:0:6:0: [sdj]  ASC=0x4 ASCQ=0x2
Oct 31 12:06:59 Servidor kernel: sd 8:0:6:0: [sdj] CDB: cdb[0]=0x28: 28 00 01 a3 7f df 00 00 10 00
Oct 31 12:06:59 Servidor kernel: end_request: I/O error, dev sdj, sector 27492319
Oct 31 12:06:59 Servidor kernel: sd 9:0:2:0: [sdn] Device not ready
Oct 31 12:06:59 Servidor kernel: sd 9:0:2:0: [sdn]  Result: hostbyte=0x00 driverbyte=0x08
Oct 31 12:06:59 Servidor kernel: sd 9:0:2:0: [sdn]  Sense Key : 0x2 [current] 
Oct 31 12:06:59 Servidor kernel: sd 9:0:2:0: [sdn]  ASC=0x4 ASCQ=0x2
Oct 31 12:06:59 Servidor kernel: sd 9:0:2:0: [sdn] CDB: cdb[0]=0x28: 28 00 01 a3 7f e0 00 00 10 00

 

When I tried to run reiserfsck on a sleepy driver, the error was trigged. :(

 

I hate to see others having the same issue as I have, but at least this seems to be more evidence that something is fundamentally broken with Linux Kernel 3.1.0, unRAID 5.0b13, LSI controller cards, and spindown/spinup.

Link to comment

I've been waiting to add 3TB drive for some time so I took chance and upgraded from 4.7 to 5.13 this weekend.

 

Basically I backed up my 4.7 usb, wiped the stick and loaded b13. I also swapped my Adaptec card for a Supermicro SASLP-MV8 and ran the permissions utility.

 

Things went well so I pushed on. I then faked a failed drive so I could do swap parity and replace a 1TB with a 3TB. I checked on it this morning it and the parity was successfully copied. The old parity 2TB drive was 35% into the build process.

 

I just remoted to check on my system and noticed the console is no longer responding. The unRaid system still responds to pings but that's about it. I tried to telnet and it refuses connection.

 

Since I have not really lost any drives I'm not too worried. Worst case is the parity drive needs to be rebuild.

 

So my options are I wait and hope it returns or force a reboot. Any suggestions or ideas? How much longer should I wait? What should I not do after a reboot? While my data is safe I still would like to keep this process as short as possible.

 

I also noticed some weirdness with permissions. Like I can access top level share but get permission denied at lower levels. If I set read to account XBMC at MOVIE share but no access at /disk1 though /diskxx shouldn't I be able to  read the folders and files under the MOVIE share? BTW reboot not correct the issue.

 

Thanks,

DB

 

 

Link to comment

My test server has now finished with 2*2TB preclearing. Preclearing went down nicely, and took approx. 24 hrs.

 

One odd observation I noticed was, that my 400GB drive is reported to be 756TB! (I wish  :P ).

 

I have attached syslog, preclear log and screendump displaying the wrong size.

 

All drives are attached to an ESXi 5.0 hardware-passthrough mapped M1015 reflashed to the P11 firmware.

 

I will now go back to test b12a on the test setup due to the spindown issue  :-\ , unless there is something specific somebody wants me to test with b13.

syslog-duringpreclear.zip

preclear_2x2TB_b13_M1015.zip

400GB_ReportedWrong.png.fcf0e1554db656d7a243d6f0210638b0.png

Link to comment
So my options are I wait and hope it returns or force a reboot. Any suggestions or ideas? How much longer should I wait? What should I not do after a reboot? While my data is safe I still would like to keep this process as short as possible.

If I saw hard drive activity on the front, I would not do a reboot until all the led's stopped showing activity.

Link to comment

If I saw hard drive activity on the front, I would not do a reboot until all the led's stopped showing activity.

 

No activity lights so I rebooted.

 

I did have the clean powerdown script installed so I did manage to get something of a log file. It's a biggie. 1.52GB to be exact. I had to use UltraEdit to open it. The entire log is basically the one restart and stripe read errors.

 

Oct 31 04:40:01 Tower syslogd 1.4.1: restart

Oct 31 04:40:01 Tower kernel: md: disk8 read error

Oct 31 04:40:01 Tower kernel: handle_stripe read error: 246879280/8, count: 1

Oct 31 04:40:01 Tower kernel: md: disk8 read error

Oct 31 04:40:01 Tower kernel: handle_stripe read error: 246879288/8, count: 1

Oct 31 04:40:01 Tower kernel: md: disk8 read error

Oct 31 04:40:01 Tower kernel: handle_stripe read error: 246879296/8, count: 1

Oct 31 04:40:01 Tower kernel: md: disk8 read error

Oct 31 04:40:01 Tower kernel: handle_stripe read error: 246879304/8, count: 1

Oct 31 04:40:01 Tower kernel: md: disk8 read error

Oct 31 04:40:01 Tower kernel: handle_stripe read error: 246879312/8, count: 1

 

The reboot brought the array up. Parity drive has green ball and array is started but the old parity drive says unformatted. I was expecting the data rebuild to take care of the formating. Not so sure what happened there. Can't say this is a beta issue or just bad luck. This morning the drive did say rebuilding.

 

Parity ST33000651AS (sdi) 2930266532 35°C 3 TB - 410 460 0

 

Hitachi HDS722020ALA330 (sdb) 1953514552 31°C 2 TB Unformatted 67 2 0

 

Disk 8 looks fine. I was able to browse the file shares and open the files.

 

Disk 8 WDC WD1001FALS (sdm) 976762552 32°C 1 TB 67.65 GB 164 46 0

Last SMART Test:  Completed without errors

 

Next steps?? I was thinking of removing the parity, adding back the 1TB drive, format the 2TB drive and copy the data from the old drive to the 2TB drive. Sound like a plan? Is there a special process for adding the parity drive back?

 

Attached is the latest log file.

 

Thanks,

DB

syslog-2011-10-31.zip

Link to comment

So the saga continues..I shut down array and unassigned the unformatted drive. The console said I can start array and it will preserve my data so I started it.

 

I then go to /mnt/disk1. This is the unformatted disk I just unassigned. I can browse my files so I think the parity might be good after all. I then stop the array and re-add my missing disk. It advises it will rebuild the disk so I start the array once again.

 

I get a Data-Rebuild in progress from the admin console.

 

Total size: 2 TB

Current position: 43.01 GB (2%)

Estimated speed: 73.87 MB/sec

Estimated finish: 442 minutes

 

I had to disable the monthly parity check to avoid that starting while rebuild is in progress. Hopefully things will look better tomorrow.

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.