unRAID Server release 4.5-beta8 available

[RobJ]

when I restart smb doesn't start. I can telnet in and start manually and it works fine.

This *may* be related to the Samba issue reported earlier, mentioned in the v4.5-beta6 summary, which includes 2 links to related threads.

 

Syslog attached.

My comments from here on are off-topic, not specific to the beta release.

 

You have a Gigabyte board and 3 drives with the Gigabyte HPA, including the parity drive.  Although not a problem now, you will have to deal with it if you add another 1.5tb drive, and at least be careful with how you connect your drives if you add even larger drives.  See the UnRAID Topical Index, HPA section.

 

Your Seagate 1.5TB parity drive needs a firmware update, see the following note from your syslog.  It currently has SD17.  Two of your Seagate 1TB drives have firmware SD15, and I would check for a firmware update for them too, but it is not as important as updating your 1.5TB Seagate.

ata6.00: HPA detected: current 2930275055, native 2930277168
ata6.00: ATA-8: ST31500341AS, SD17, max UDMA/133
ata6.00: 2930275055 sectors, multi 16: LBA48 NCQ (not used)
ata6.00: WARNING: device requires firmware update to be fully functional.
ata6.00:          contact the vendor or visit http://ata.wiki.kernel.org.

[lewcass]

I'll look into this.

 

Thanks much. I know Realtek has been updating the drivers for Vista very frequently, apparently they have trouble optimizing them. Suspect this may also be the case for Linux and this kernel may just have a bad one. Appreciate anything you can do, so I can take advantage of the latest improvements in unRAID without having to purchase a new NIC. (Leaves more funds for me to upgrade to Plus.   )

 

What motherboard are you using.  From the syslog it appears you have a Marvell NIC, not Realtek.

 

DUH!   Asus A8R32MVP Deluxe which does have Marvell network controllers. I have a Realtek in my workstation's motherboard!

 

Got myself confused.

 

Sorry to be an idiot customer who complicates a problem more than necessary!

 

No worries.  That m/b has two Marvell NIC's.  I guess you have one of them disabled - can you switch over to the other one as a test?

 

Yes. The secondary one is PCI not PCIe, but I'll go ahead and test it with 4.5-b8 if that will provide you with a useful data point...

 

Well, the secondary NIC (888001 PCI Gigabit) would not negotiate a gigabit connection with beta 6 but it does with beta 8!

 

Of course I'd rather be using the primary NIC (88E8053 PCIe Gigabit) which displays the opposite behavior. (Edit: I especially don't want to use the secondary NIC since I discovered that WOL does not seem to work with it.) 

 

Nothing like complicating a problem is there?

 

Looks like I've got this solved.  Though I'll have to test a while yet to make certain that gigabit works consistently.

 

I enabled both NICs in the BIOS and for some reason this apparently is allowing the primary PCIe NIC to negotiate a gigabit connection, while it would not do so with the secondary NIC disabled.

 

This was not the case with 4.5-beta 6 where I had been successfully using gigabit with only the primary enabled.  #$%&@ computers.

 

Anyway, thanks Tom for looking into this to the extent you did.

 

 

[redbear]

the great news... my write speed into my array went from 15 MB/sec with beta7 to 30 MB/sec with beta 8...

 

on the other hand i got write speeds of 100 MB/sec with the same hardware using win 2008 r2.

[jdubs]

on the other hand i got write speeds of 100 MB/sec with the same hardware using win 2008 r2.

 

Parity protected?

[michael123]

on the other hand i got write speeds of 100 MB/sec with the same hardware using win 2008 r2.

 

Parity protected?

 

1) Are you writing direct to disk or through user share?

2) Which hardware are you using?

 

I had boost as well, but from merely 8-9 MB/s to 19-20 MB/sec

[Guzzi]

on the other hand i got write speeds of 100 MB/sec with the same hardware using win 2008 r2.

To my experience those rates only apply when 2008r2 writes to the RAM cache - I never got those rates continously (with standard HW). If talking about parityprotected arrays, take into consideration differences between raid levels used.

Back to topic: I also confirm boost of transfer speed by almost 100% plus working datadrive #19 - very nice and thank you.

[redbear]

the great news... my write speed into my array went from 15 MB/sec with beta7 to 30 MB/sec with beta 8...

 

on the other hand i got write speeds of 100 MB/sec with the same hardware using win 2008 r2.

 

Sorry,for the lack of details. I hope the following helps clarify.

 

For my test I wrote the same 4 GB file to and copied it from:

 

unraid pro  4.5-beta8: wrote to a  user share which was comprised of 3 disks + the parity disk. (I know the file just writes to one disk + the parity disk, just letting you know the setup)

 

2k8: wrote to a share on a software raid 5 array comprised of 3 disks (each disk with one partition taking up the entire drive)

 

I run an Abit 9 Pro using all 9 internal sata ports + a Promise tx-300 using 3 of its ports. All drives are Western Digital Caviar Green 1TB's, with the exception of one lone Hitachi E7K500.

 

Background: I picked up 4 new 1TB drives so I thought it would be a good time to have a Windows Home Server, 2k8, Unraid, Ubuntu, OpenSolaris shootout. I've been really itching to consolidate servers/services since moving and I wanted something that I could easily run vm's on. With the release of beta8, I thought let's do it.

[prostuff1]

2k8: wrote to a share on a software raid 5 array comprised of 3 disks (each disk with one partition taking up the entire drive)

 

The software RAID5 part is were the difference will mainly be.  unRAID does not strip data like a RAID5 so it will only ever be as fast as what the drive can do.  Also, try writing directly to the disk in the unRAID environment to see if the speed is any better... it usually is.

[boof]

Background: I picked up 4 new 1TB drives so I thought it would be a good time to have a Windows Home Server, 2k8, Unraid, Ubuntu, OpenSolaris shootout. I've been really itching to consolidate servers/services since moving and I wanted something that I could easily run vm's on. With the release of beta8, I thought let's do it.

 

All of these would beat unraid in raw throughput stakes assuming (WHS excepted) you're looking at software RAID or RAIDZ with them (although a cache drive in unraid would muddy the waters with WHS at least).

 

unraid is principally a completely different attitude to parity protection than any of those solutions - its these differences from 'normal' RAID arrays that are amongst the most appealing things about unraid.

 

[speeding_ant]

Cache drive working now, but when I restart smb doesn't start. I can telnet in and start manually and it works fine. Syslog attached.

 

Is this repeatable, that is, do you see this every time you reboot the server?  If so, please hook up a monitor, re-boot, and tell me if you see any error messages output near the end of the boot process (there are certain messages that only show up on the console & not in the system log).

 

It's odd. More often than not it does repeat itself. Just after I upgraded the firmware on the 1.5Tb Seagate last night, SMB started straight away, then restarted server again and it didn't start SMB... And then you start manually and all is well. Nothing else out of the ordinary pops up when you look at output on monitor.  Odd! Also having weird permissions errors. I can create a folder but can't rename it on certain directories. Here's an updated syslog.

[WeeboTech]

Where is the md driver source and .config for this release ?

 

[tdoris]

I'm still seeing very high shfs cpu usage with this release compared with 4.4.2.  If I have 2 HD streams active recording and playing back (sagetv live tv), I'll generally see stuttering at times in one.  I don't believe this should be happening with the CPU I'm using (AMD Athlon Dual Core Processor 4850e) with an ASUS M2R32-MVP motherboard.  I posted a syslog in the Beta 7 thread, but I'll attach another one. 

[RobJ]

I'm still seeing very high shfs cpu usage with this release compared with 4.4.2.  If I have 2 HD streams active recording and playing back (sagetv live tv), I'll generally see stuttering at times in one.  I don't believe this should be happening with the CPU I'm using (AMD Athlon Dual Core Processor 4850e) with an ASUS M2R32-MVP motherboard.  I posted a syslog in the Beta 7 thread, but I'll attach another one. 

 

Syslog looks fine, nothing problematic is evident.  Sorry but I must have missed your previous syslog.

 

Off-topic, but hard not to be envious of your equipment!  When you have a chance, I hope you'll report your read/write performance numbers on that Cache drive, in a different thread.

[wholly]

As I thought I understood things, I thought unraid ran without any swap.  After a couple parity checks and adding two drives (currently being cleared) I'm seeing 60+% of my CPU being used by kswapd0. 

 

Is this sane?  (I am running on 1G of ram on beta8)

[limetech]

Where is the md driver source and .config for this release ?

 

 

ls -Ra /usr/src/linux/

[limetech]

As I thought I understood things, I thought unraid ran without any swap.  After a couple parity checks and adding two drives (currently being cleared) I'm seeing 60+% of my CPU being used by kswapd0. 

 

Is this sane?  (I am running on 1G of ram on beta8)

 

From this posting on kerneltrap:

 

"kswapd also reclaims pagecache, not just anonymous memory. It runs

in response to memory pressure and if it wasn't around, then all

your apps requesting memory would have to do basically the same

amount of work themselves."

 

In other words, it's a garbage collector 

[limetech]

I'm still seeing very high shfs cpu usage with this release compared with 4.4.2.  If I have 2 HD streams active recording and playing back (sagetv live tv), I'll generally see stuttering at times in one.  I don't believe this should be happening with the CPU I'm using (AMD Athlon Dual Core Processor 4850e) with an ASUS M2R32-MVP motherboard.  I posted a syslog in the Beta 7 thread, but I'll attach another one. 

 

High CPU usage by shfs is to be expected.  This is the FUSE-based process that is implementing the 'share' file system.  FUSE tends to be CPU-intensive.

 

I'm more interested in the 'stuttering' you are seeing.  What do you mean by "I'll generally see stuttering at times in one"?  You have one write stream (the one recording) and one read stream (the one you're watching), did you mean to say, "I'll generally see stuttering at times in the read stream"?  Two more questions:

1. Are these streams accessing the same data disk?

2. Have you changed 'md_num_stripes' and/or 'md_write_limit' from their default values?

[Marty068]

Hello,

 

I'm new on the forum. I found unRAID a great product 

I just upgraded my unRAID server from 4.4.2 to 4.5-beta8 to solve my problem of stand-by (S3).

After the upgrade I noticied the lost of connectivity with my Nas.

I get the following error in my console:

e100:eth0:e100_request_firmware:failed to load firmware "e100/d101m_ucode.bin" -2

My nic is an intel 100 (pro?), the led from the NIC are off no more activity

Seems to be a problem with the kernel (I have made a search with Google)

How this can be fixed with 4.5-beta8?

My knowledge with Linux are very basic.

My Nas is now running again the 4.4.2 and connectivity is back.

 

Thanks in advance

Marty068

[limetech]

Hello,

 

I'm new on the forum. I found unRAID a great product 

I just upgraded my unRAID server from 4.4.2 to 4.5-beta8 to solve my problem of stand-by (S3).

After the upgrade I noticied the lost of connectivity with my Nas.

I get the following error in my console:

e100:eth0:e100_request_firmware:failed to load firmware "e100/d101m_ucode.bin" -2

My nic is an intel 100 (pro?), the led from the NIC are off no more activity

Seems to be a problem with the kernel (I have made a search with Google)

How this can be fixed with 4.5-beta8?

My knowledge with Linux are very basic.

My Nas is now running again the 4.4.2 and connectivity is back.

 

Thanks in advance

Marty068

 

Every once in a while a kernel developer makes a change that seems like a good idea at the time, but has unintended consequences (a.k.a. brain fart).  This is apparently one of those occasions.  From a thread on kerneltrap, 2.6.29 included a change to the way e100 firmware is loaded.

 

I'll put the fix to this in -beta9.

[Marty068]

Every once in a while a kernel developer makes a change that seems like a good idea at the time, but has unintended consequences (a.k.a. brain fart).  This is apparently one of those occasions.  From a thread on kerneltrap, 2.6.29 included a change to the way e100 firmware is loaded.

 

I'll put the fix to this in -beta9.

 

Thanks!

Waiting for next release 

 

[tdoris]

I'm still seeing very high shfs cpu usage with this release compared with 4.4.2.  If I have 2 HD streams active recording and playing back (sagetv live tv), I'll generally see stuttering at times in one.  I don't believe this should be happening with the CPU I'm using (AMD Athlon Dual Core Processor 4850e) with an ASUS M2R32-MVP motherboard.  I posted a syslog in the Beta 7 thread, but I'll attach another one. 

 

High CPU usage by shfs is to be expected.  This is the FUSE-based process that is implementing the 'share' file system.  FUSE tends to be CPU-intensive.

 

I'm more interested in the 'stuttering' you are seeing.  What do you mean by "I'll generally see stuttering at times in one"?  You have one write stream (the one recording) and one read stream (the one you're watching), did you mean to say, "I'll generally see stuttering at times in the read stream"?  Two more questions:

1. Are these streams accessing the same data disk?

2. Have you changed 'md_num_stripes' and/or 'md_write_limit' from their default values?

 

Generally what cause this is as follows:

1) I'll start using livetv on both my hd100 and hd200 receiver boxes on two different HD (1080i) channels

2) This causes sagetv to start 2 recordings (write streams) and 2 playback (read streams) to my cache disk

3) Periodically (usually when CPU utilization reported by top exceeds 100%) I'll see stutter in 1 of the playback (read) streams.

 

This doesn't happen in 4.4.2, but does happen in all of the 4.5-betas I've tried.  CPU utilization seems to be a factor of 3-4 times higher than in 4.4.2. 

 

I've tried playing a bit to with the md_num_stripes and md_write_limit, but it doesn't seem to be making much a difference.  CPU utilization seem less in beta8 then beta7 (where I couldn't adjust the md parameters).

 

If this is just the way 4.5 is going to be, has limetech been looking at any newer motherboards to be officially supported?  I'd like to get something that supports 4 PCI-e x8 slots like the Asus M4A79 Deluxe, but if there is something else in the pipe I'd definitely consider getting it instead.

[screwonbudnik20]

My speeds writing to  user shares are vastly improved was 20-23MB now 60-65 sustained 100+ gig transfers, never tried going direct to disk as I have cache drive installed. No problems thus far streaming BR is also flawless. I wish I had the capabilty to try multiple streams at once but I don't have the equipment.

[cwr]

has anyone noticed disk spindown not working on their unraid server?

 

i'm running a supermicro c2see board and tried different spindown times (15m, 45m, 1h, 2h) and none seem to work. all of this worked fine in beta7

[fitbrit]

has anyone noticed disk spindown not working on their unraid server?

 

i'm running a supermicro c2see board and tried different spindown times (15m, 45m, 1h, 2h) and none seem to work. all of this worked fine in beta7

 

My spin-down is working perfectly with this beta, using two Promise TX-4 PCI cards, an eSATA Sil PCIe card and onboard nVidia SATA ports.

[michael123]

Hi Tom

 

Can we get suggested tuning for various configurations,

for those who have quit powerful CPU and 4GB RAM?

 

-- Michael