Re: ARC-1200 in SAFE33 mode vs M1015


Recommended Posts

When I said do the speed test it was on the ARC1200 array drives (before formatting for unRAID) In any case movement of the controller improved your speed. Your speeds are what I would expect. I would however suggest you purchase two 7200RPM drives for the SAFE33 volume.  Although members will say you are limited to to your slowest drive, if you use more then one drive or do allot of random I/O as with bittorrent, the extra speed helps. I know it's a cost thing, but that's my experience.

Link to comment

I switched the ARC1200 to the other PCIe x4 port. I now get parity sync speeds of 85-90 MB/s. Can't believe it makes such a difference. Let's see what the Parity check does later on.

 

 

Parity check vs parity sync will be slightly different, but close to one another. With sync, you are writing and the wb cache is helping.

With sync, you are reading and the cache does not help as much.

Link to comment

When I said do the speed test it was on the ARC1200 array drives (before formatting for unRAID) In any case movement of the controller improved your speed. Your speeds are what I would expect. I would however suggest you purchase two 7200RPM drives for the SAFE33 volume.  Although members will say you are limited to to your slowest drive, if you use more then one drive or do allot of random I/O as with bittorrent, the extra speed helps. I know it's a cost thing, but that's my experience.

I narrowed my choice for 7200rmp 2TB drives down to:

Seagate Barracuda 7200.14 ST2000DM001, 2TB

Hitachi Deskstar 7K3000 HDS723020BLA642, 2TB

Hitachi Deskstar 7K2000, 2TB

Seagate Barracuda STBD2000201, 2TB

 

They are within 10 Euros here in Holland, which would your choice? The 7K3000 is my fovorite with the 7K2000 2nd place, but don't ask me why.

Link to comment

When I said do the speed test it was on the ARC1200 array drives (before formatting for unRAID) In any case movement of the controller improved your speed. Your speeds are what I would expect. I would however suggest you purchase two 7200RPM drives for the SAFE33 volume.  Although members will say you are limited to to your slowest drive, if you use more then one drive or do allot of random I/O as with bittorrent, the extra speed helps. I know it's a cost thing, but that's my experience.

I narrowed my choice for 7200rmp 2TB drives down to:

Seagate Barracuda 7200.14 ST2000DM001, 2TB

Hitachi Deskstar 7K3000 HDS723020BLA642, 2TB

Hitachi Deskstar 7K2000, 2TB

Seagate Barracuda STBD2000201, 2TB

 

They are within 10 Euros here in Holland, which would your choice? The 7K3000 is my fovorite with the 7K2000 2nd place, but don't ask me why.

 

 

Speaking frankly, I don't have a preference at this time. I have not added or researched drives in a while.

I remember the Hitachi's being reported better at Random I/O. How true it is I'm not sure, the benchmarks showed that also.

Go with your favorite or wait for other recommendations. Perhas Raj has a favorite too.

Link to comment

I switched the ARC1200 to the other PCIe x4 port. I now get parity sync speeds of 85-90 MB/s. Can't believe it makes such a difference. Let's see what the Parity check does later on.

 

 

Parity check vs parity sync will be slightly different, but close to one another. With sync, you are writing and the wb cache is helping.

With sync, you are reading and the cache does not help as much.

 

Parity sync speed started a 95 MBs then gradually dropped to 55-60 at 1 TB (size of the largest array-disk). At the 1 TB point speed went up to 90-95 MBs and now at 70% (2.1 TB) is at 88 MBs. Only the 3 TB RAID0 volume (parity) is now written to.

Link to comment

I switched the ARC1200 to the other PCIe x4 port. I now get parity sync speeds of 85-90 MB/s. Can't believe it makes such a difference. Let's see what the Parity check does later on.

 

 

Parity check vs parity sync will be slightly different, but close to one another. With sync, you are writing and the wb cache is helping.

With sync, you are reading and the cache does not help as much.

 

Parity sync speed started a 95 MBs then gradually dropped to 55-60 at 1 TB (size of the largest array-disk). At the 1 TB point speed went up to 90-95 MBs and now at 70% (2.1 TB) is at 88 MBs. Only the 3 TB RAID0 volume (parity) is now written to.

 

 

 

This is normal.

Link to comment

Hope it stays stable this way.

 

Regarding harddisk choice: I won't go for a very expensive RAID, 24/7 disk. They cost double!

 

 

It should be fine,

Also you don't need the data center 24/7 raid drives. unRAID usually works fine with regular drives.

 

I think the best "regular" drives are: the ones you find at a good price and warranty length within your comfort.  For example: I bought a new 2TB drive - the Seagate was on sale for $119 or so, and the WD Red was $139.  I bought the WD Red as it has a 3 year warranty vs. Seagate now having only 1 year warranty on "regular" drives.  Well worth the extra $20 in my opinion.

Link to comment

Hope it stays stable this way.

 

Regarding harddisk choice: I won't go for a very expensive RAID, 24/7 disk. They cost double!

 

 

It should be fine,

Also you don't need the data center 24/7 raid drives. unRAID usually works fine with regular drives.

 

I think the best "regular" drives are: the ones you find at a good price and warranty length within your comfort.  For example: I bought a new 2TB drive - the Seagate was on sale for $119 or so, and the WD Red was $139.  I bought the WD Red as it has a 3 year warranty vs. Seagate now having only 1 year warranty on "regular" drives.  Well worth the extra $20 in my opinion.

 

Thank you marcus, that sure is a valid point. I don't exactly know how the situation is here in Europe/Holland. And with all those companies joining. I don't know the WD REDs, but I see them at far higher prices than you mention. Are they 7200 rpm drives, as that is what I am looking for. My experience with WD20EARS was rather bad, 4 died within 10 months. WD did send me replacements in advance, but it was a rather costly affair because the broken ones had to be shipped to Germany, WD has no location in Holland. So I rather avoid that.

Link to comment

42MB/s is not what I would expect from RAID0 even on x1 lane.

 

 

You could try moving it to other slots to see what the deal is.

On my card, I enabled write back cache.

 

 

The best test is to use the writeread10gb script on these raw drives to determine maximum speed before you build the array.

 

 

I.e. with the 3 drives and emhttp down, format each drive with reiserfs manually.

After that  use the writeread10gb script on each drive.

That will be the maximum you could possibly achieve on each drive. From there you can factor in all the other issues.

I can say that if you are doing a parity operation (read/write/sync, etc, etc) and write data to the cache drive, there will be a performance penalty and it could be significant. So I do those two operations separately.

 

 

If you have a UPS you can enable the write back cache on the controller in the firmware before doing anything else to see if it helps.

 

 

I have 15 data drives to 1 ARC1200 on a x1 slot, I get from 75-80 at the start dropping to 55k near the end.

Jul 14 21:22:10 atlas kernel: md: sync done. time=18244sec rate=53538K/sec

 

 

I have 2 seagate 1.5tb 7200RPM 32MB cache drives as my raid0/raid1 hybrid.

If you are using 5400rpm drives for yours, that could be an issue.

 

I did the writeread10gb test.

 

Both disks showed write speeds averages of 36-40 MBs

The RAID1 cache volume on the ARC1200 showed an average of 55-60 MBs

 

Reading test gives an erroneous value of 1.4 GBs

Link to comment

Those speeds are good.

 

 

With some tunings they might get better depending on how much memory you have.

I set my tunables to these values (2x previous values)

 

Tunable (md_num_stripes): user-set
Tunable (md_write_limit): user-set
Tunable (md_sync_window): user-set

 

I also set my kernel tunables to

 

 

 

sysctl vm.vfs_cache_pressure=10

 

sysctl vm.swappiness=100

 

sysctl vm.dirty_ratio=20 (you can set it higher as an experiment).

sysctl -w vm.min_free_kbytes=8192

 

This helps by caching more in ram (at the potential risk of loss due to sudden loss of power).

If you are using a UPS, then this is not that much of a risk.

 

 

What this lets me do is burst at up to 60MB/s to the array.

Which is good for small files. I get it onto the server faster and let the unix cache flush it out to disk.

 

 

Since I use my unRAID server as a general file server and not just a media server, it's important for me to get good speed on small files.  Especially since I NFS mount all my source from the unRAID server.

Link to comment

Those speeds are good.

 

 

With some tunings they might get better depending on how much memory you have.

I set my tunables to these values (2x previous values)

 

Tunable (md_num_stripes): user-set
Tunable (md_write_limit): user-set
Tunable (md_sync_window): user-set

 

I also set my kernel tunables to

 

 

 

sysctl vm.vfs_cache_pressure=10

 

sysctl vm.swappiness=100

 

sysctl vm.dirty_ratio=20 (you can set it higher as an experiment).

sysctl -w vm.min_free_kbytes=8192

 

This helps by caching more in ram (at the potential risk of loss due to sudden loss of power).

If you are using a UPS, then this is not that much of a risk.

 

 

What this lets me do is burst at up to 60MB/s to the array.

Which is good for small files. I get it onto the server faster and let the unix cache flush it out to disk.

 

 

Since I use my unRAID server as a general file server and not just a media server, it's important for me to get good speed on small files.  Especially since I NFS mount all my source from the unRAID server.

 

 

How do I set these kernel tunables?

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.