jademonkee

New HP MicroServer gen8 w/ 2x 4TB - 33MB/s building parity normal?

Recommended Posts

'ullo!

It's been a crazy few days! I was just about to buy a Synology DS916+ when a friend told me about the HP MicroServers and unRAID (I'd not heard of unRAID), and after a few days research and decision making, I decided that unRAID on a MicroServer was better in almost every way (the only exception being power consumption) than the Synology NAS, and found a good deal on a new gen8 MicroServer w/ 16GB RAM, and pulled the trigger.

I've now set it up (firmware updates, enabling AHCI), and am running unRAID with 2x 4TB Seagate NAS drives in bays 1 & 2, and the parity is now syncing.

Being so new to this, I of course now have some questions (most of which have been answered by the handy search bar).

The one on my mind right now is about speed: the parity is building at ~33MB/sec, which seems slow (1 day, 53 minutes remaining after 9 hours so far).

Is 33MB/sec perfectly normal for building the parity drive, or is there some setting I can change to speed it up?

Further, is 33MB/s what I can expect for writing speeds until I save up for a cache drive (I was thinking it'd be around 60MB/s)?

Many thanks.

Share this post


Link to post
Share on other sites

No, it's not normal, with only 4 disks it should sync at the disks max speed, for those disks should start at ~160Mb/s and then steadily decrease as it goes through the inner sectors to about 80Mb/s.

Share this post


Link to post
Share on other sites

It's 2x disks, but that's upsetting news.

Could it perhaps be that the disks were damaged in transit?

Are there any tests I can run?

Share this post


Link to post
Share on other sites

I have a HP N40L Microserver and my 4TB WD RED parity check normally completes within 10 hours and at around ~110MB/s

 

Transfrer speeds were around ~60MB/s direct to disk but with SSD cache I get ~110MB/s

Share this post


Link to post
Share on other sites

Hrmmmm... thanks for the info. Extra disappointing to see these speeds on my end, then.

Hopefully a solution will be found soon (and hopefully one that doesn't require me returning the drives).

Share this post


Link to post
Share on other sites

Diagnostics look normal, don't see anything wrong, you can run diskspeed to check if speeds are normal, you have to wait for the parity sync to end (or cancel it) to get accurate results.

 

Share this post


Link to post
Share on other sites

If I stop the parity, will it pick up where it left off when I start it again? Or will it start from scratch?

Share this post


Link to post
Share on other sites

If I stop the parity, will it pick up where it left off when I start it again? Or will it start from scratch?

 

It always starts from the beginning.

Share this post


Link to post
Share on other sites

I decided to stop the parity sync, and find out any drive problems before going further.

I copied the drive performance script to the system USB drive, and executed it from the root account over SSH.

It returned this error:

"Error: There are invalid drives in the array"

Is this because of the unfinished parity sync?

Share this post


Link to post
Share on other sites

Yes, forgot about that, you can do a new config and assign both disks as data devices to run the test, in the end you'll need to do another new config and re-assign them as they are now to begin the parity sync.

Share this post


Link to post
Share on other sites

Ok, just ran the test and got the following results:

/dev/sdb (Disk 2): 173 MB/sec avg

/dev/sdc (Disk 1): 159 MB/sec avg

 

sdb is the previous parity drive.

 

FWIW, I just copied a 400 MB collection of lossless audio files over to sdc (containing my empty media share) at about 95-100MB/s.

 

Any idea what else may be causing the slow speeds?

 

I shall try rebuilding the parity now, and report back.

 

EDIT: Same speeds building the parity as before :( I guess I'll just have to wait the day until it completes, then test to see what write speeds are. Maybe there's something funny going on with the controller causing slow speeds when reading/writing at the same time. I dunno. I just hope it doesn't mean that writes over the network will be 35MB/s...

Share this post


Link to post
Share on other sites

Any idea what else may be causing the slow speeds?

 

Not really, I remember there is a similar issue with some Hitachi disks using AHCI, they are very slow syncing parity but then perform normally, including parity checks, maybe these Seagtes have the same issue, I would let the parity sync finish and then test read/write performance and also parity check speed, if those are normal then it's probably it.

 

Note also that with only 2 disks unRAID works almost like RAID1, so write performance should be very good, usually limited by gigabit, you will notice a significant slowdown if/when you add more disks, normal parity protected write speed is 45 to 60MB/s.

Share this post


Link to post
Share on other sites

Thanks again for the help and the info.

The friend who recommended me unRAID warned me of the slow speeds, so I'm expecting them.

What I didn't know was that it's faster with just the 1+1 disks. I was looking to add two SSDs for cache some time in the near future, but it's good to know that I won't have to worry about that until I add in another storage disk.

Thanks again.

Share this post


Link to post
Share on other sites

You may need to set the array to JBOD/AHCI. Also, the 5th SATA port runs at 3Gbit/sec IIRC.

On mine, I set it up as AHCI and not as a RAID IIRC. 

Share this post


Link to post
Share on other sites

In BIOS, you mean? I have set the card to AHCI (Press F9 during boot -> System Options -> SATA Controller Options -> Embedded SATA Configuration -> Enable SATA AHCI Support)

Or did you mean in an unRAID setting (or somewhere else entirely)?

Share this post


Link to post
Share on other sites

Hang on a sec, I actually had the same thing once not so long ago where I left my browser open and it showed a slower speed, but soon as I closed it and just left it carried on as normal and finished around 10 hour mark.

 

I would suggest running parity and leave it overnight dont bother keep going back and checking and make sure the browser is closed. If after say 12 hours you check and have something like this

Last checked on Sun 03 Jul 2016 12:49:28 AM BST (four days ago), finding 0 errors.
Duration: 10 hours, 24 minutes, 26 seconds. Average speed: 106.8 MB/s

 

I dont think you have anything to worry about. Failing that I would google gen8 bios settings and make sure everything is set right.

 

***update***

Just found my old thread https://lime-technology.com/forum/index.php?topic=49175.msg471822#msg471822 and believe it was a combination of both, leaving browser open and failing disk.

 

Hope this helps.

Share this post


Link to post
Share on other sites

Thanks for the info, but I've had the browser closed for the majority of the time, and it doesn't look to change the speed at all.

I think it might be something in the Seagate NAS drive's firmware. To quote the marketing material: "supports error correction via customised error recovery controls" so there might be something going on there.

There's still 16 hours left for the parity build, but I shall run some speed tests once it's done and report back.

Share this post


Link to post
Share on other sites

Check your system log and see if one or more of your drives are connecting as if they were an IDE drive rather than at SATA speeds.  I had the same problem when I connected my HGST NAS drives to my N54L and I could see one drive at IDE speeds.  With standard desktop Seagate drives they connected at full SATA speeds and all slots full of drives.  I thought it was an incompatibility with HGST but now it may just be the NAS function like you were speculating about.

 

 

Here were my symptoms:

I had a IBM M1015 controller installed in my N54L and when I switched the HGST NAS drives all to the M1015 the speeds returned to normal.  But no matter what drive I attached to the built in controller on the N54L (and in any slot it didn't matter which) the last drive was always at IDE speeds.  I.E. with one drive that drive was IDE speed. With two drives on the N54L controller and the drive in slot 2, 3 or 4 (didn't matter which slot or drive) it would connect at IDE speeds and the drive in slot 1 would be at SATA speeds.  Swap drives around and slots and still last drive at IDE speeds.  Add a third drive and it became the IDE Speed drive and the other two would be at full SATA speeds.  Etc...

Share this post


Link to post
Share on other sites

Check your system log and see if one or more of your drives are connecting as if they were an IDE drive rather than at SATA speeds.  I had the same problem when I connected my HGST NAS drives to my N54L and I could see one drive at IDE speeds.  With standard desktop Seagate drives they connected at full SATA speeds and all slots full of drives.  I thought it was an incompatibility with HGST but now it may just be the NAS function like you were speculating about.

 

 

Here were my symptoms:

I had a IBM M1015 controller installed in my N54L and when I switched the HGST NAS drives all to the M1015 the speeds returned to normal.  But no matter what drive I attached to the built in controller on the N54L (and in any slot it didn't matter which) the last drive was always at IDE speeds.  I.E. with one drive that drive was IDE speed. With two drives on the N54L controller and the drive in slot 2, 3 or 4 (didn't matter which slot or drive) it would connect at IDE speeds and the drive in slot 1 would be at SATA speeds.  Swap drives around and slots and still last drive at IDE speeds.  Add a third drive and it became the IDE Speed drive and the other two would be at full SATA speeds.  Etc...

 

Was there ever a solution to this? Or did you have to return the NAS drives?

Also, I'm not sure what to search for in the logs? I searched for "ide" but found no instances.

I'm not sure that's the problem, anyway, as I ran diskspeed and got the following results:

/dev/sdb (Disk 2): 173 MB/sec avg

/dev/sdc (Disk 1): 159 MB/sec avg

 

I'm currently copying across the network at about 80MB/s (large files).

Share this post


Link to post
Share on other sites

I had a gen 8 microserver running unraid for a while and parity was much quicker. I would double check all your bios settings just in case

 

Sent from my SGP612 using Tapatalk

 

 

Share this post


Link to post
Share on other sites

I had a gen 8 microserver running unraid for a while and parity was much quicker. I would double check all your bios settings just in case

 

What should I look for? Is AHCI the only setting that should be changed from default?

 

EDIT: Have booted to BIOS.

System Options > SATA Controller Options

Embedded SATA Configuration: Enable SATA AHCI Support

Drive Write Cache: Disabled

Share this post


Link to post
Share on other sites

Sorry i cant remember the bios settings or options on the micro gen 8.

But just make sure no disk controllers are in raid mode as unraid needs direct access to the disks. But seeing as you have selected sata ahci it should be ok. Maybe my speeds were faster as i had changed the cpu from the dual core celeron to a quad xeon. The slower parity could be down to the calculations needed for parity, but im not sure. Someone else here would know better than me if parity speed needs much cpu.

 

Once you have your build running if you dont already have one i would suggest using a small ssd cache drive. That makes a big difference on general use for write speeds. I put mine where the sata dvd drive goes so i didnt have to use one of the main drive bays.

If i remember i had to make up a cable adaptor for the power socket on the dvd drive power cable.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Copyright © 2005-2017 Lime Technology, Inc. unRAID® is a registered trademark of Lime Technology, Inc.