Highpoint controller info

Status
Not open for further replies.

cyberjock

Inactive Account
Joined
Mar 25, 2012
Messages
19,526
So I've been experimenting with various Highpoint controllers. There are posts every week or 2 involving people with problems with their Highpoint controllers. In the past Highpoint controllers have been a cost effective hardware RAID. I've put together this post to help people that have the controller family below. This post is the result of 5 days of experimenting with several controllers I own. Use the information below at your own risk.

I have personally experimented with the 22xx, 23xx, 35xx and 43xx series adapters. The adapters I used in particular are the 2220, 2320, 3560 and 4320(yes, I used highpoint controllers for many years as servers for home use and small business). As of FreeNAS build 12037 and newer the entire 35xx and 43xx line of controllers should function properly. Prior to build 12037 there are driver issues I corrected via ticket 1704 for some of the 35xx and 43xx series controllers(thanks delphij!). All builds including and after 8.3.0-BETA1 include the fix for the affected controllers.

Keep in mind that FreeNAS performs most effectively when NOT part of a hardware RAID. My comments below will reflect that philosophy. I will also assume that you will be creating zpools that involve mirrors or parity for reliability.

Note: SMART data and serial numbers are NOT currently working with ANY Highpoint controllers. This is a feature that is being actively pursued by myself and William Grzybowski(developer). Serial and SMART may or may not work ever. The Highpoint CLI doesn't work "out of the box" with FreeNAS as well as other issues that can be resolved, but will be time consuming to implement. It is possible that the 22xx and 23xx will never have support for SMART and serial numbers because they are so old and the development resources needed to make these work are just not worth the effort. I will update this page with more information when I have more solid information. Sorry for not having a firm answer.

Here's a list of limitations and problems with the various families:

22xx & 23xx - These controllers do NOT mount hard drives that are not part of a RAID array as a "Legacy Disk". To be mounted as a "Legacy Disk" they MUST contain a valid partition table with at least one partition. Unfortunately ZFS partitions are NOT identified as valid partitions.

The workaround is to create a partition on the hard drive that is FAT16, FAT32 or NTFS of any size. In my case I created a 100MB partition that is FAT16. If at any time your FAT16 partition table entry is deleted or otherwise corrupted you WILL lose the ability to access your ZFS partition when connected to a Highpoint controller.

Also it should be mentioned that if a disk fails(in my case I tested this by disconnecting the power to a hard drive.. do not do this if you have data on the zpool!!!) you will get the RAID controller audible alarm. FreeNAS will not acknowledge the failed drive since the disk is not "detached" from the controller in the software sense. Presumably this is because these controllers do not support hot swap on FreeBSD. The only way you will be able to acknowledge that a disk is failed is possibly a local LED(one disk does not have any activity while the rest of the drives do), doing a 'zfs status -v' and noticing that one drive has alot of errors, audible alarm or if you reboot the machine. If you reboot the machine the status of the zpool will change to DEGRADED. Note that the zpool status will not change until you reboot, but I'd expect that the audible alarm will get your attention rather quickly. It's very loud and quite annoying. The only way I know of to silence the alarm is to shutdown/reboot the server.

Overall, if you have the audible alarm enabled that should provide sufficient warning that a drive has failed. Since hot-swap is not supported you will need to shutdown the server, insert a new hard drive then power up the server and add the new hard drive to the zpool per the FreeNAS manual.

To perform the workaround you will need to connect each hard disk to a standard SATA controller to create the partition table entry. After the entry is created you can then remove the drive and it will function correctly as a "Legacy Disk" on the Highpoint controller.

Note: If you plug in a hard drive that is blank(aka no partition table at all) it will be identified as "New" to the controller. My computer panic'd during bootup anytime a "New" hard drive was attached to my 2220 and 2320. Because of this you MUST connect the hard drive to a different controller(onboard recommended) to setup the partition.

Determine which device your hard drive is assigned to. Use the 'dmesg' command to view the devices. My devices are always adaX or daX. For the following example my hard drive was ada1:

At a console, type the following commands. If you have data on the hard drives make sure you backup the data as these commands will erase all data on the hard drives.

# dd if=/dev/zero of=/dev/ada1 bs=512 count=10k
# gpart create -s gpt ada1
# gpart add -t mbr -s 100M ada1
# newfs_msdos /dev/ada1p1

Because these Highpoint controllers do not work out of the box and require this workaround it is not recommended that you use the 22xx or 23xx series controllers in a production environment. Additionally any hard drive that is ever used on this controller in the future(spares/upgrades,etc) will require this modification before installing in the FreeNAS server if attaching the hard disk to the RocketRAID controller to work properly. Do not expect wide support and assistance if you use these controllers. You are on your own and at your own peril by following the instructions above. I will try to help if you post in this thread, but if you value your data I would not recommend you use these old controllers. Nonetheless I'm sure some people will do it anyway, and I'm writing this post to hopefully once and for all answer the questions people have about these 2 controller families.


35xx & 43xx - These controllers will mount all hard drives as a "Legacy Disk" if the RAID mode is changed to Non-RAID adapter mode. This will allow for proper operation of the hard disks with no changes as listed on the 22xx or 23xx series controllers above.

The optimal configuration is to setup the controller in "Non-RAID adapter mode"in the controller's BIOS. After this is done and a reboot performed the hard drives will be picked up as daX devices.

Also it should be mentioned that if a disk fails(in my case I tested this by disconnecting the power to a hard drive.. do not do this if you have data on the zpool!!!) you will get the RAID controller audible alarm. FreeNAS will not acknowledge the failed drive since the disk is not "detached" from the controller in the software sense. Presumably this is because these controllers do not support hot swap on FreeBSD. The only way you will be able to acknowledge that a disk is failed is possibly a local LED(one disk does not have any activity while the rest of the drives do), doing a 'zfs status -v' and noticing that one drive has alot of errors, audible alarm or if you reboot the machine. If you reboot the machine the status of the zpool will change to DEGRADED. Note that the zpool status will not change until you reboot, but I'd expect that the audible alarm will get your attention rather quickly. It's very loud and quite annoying. The only way I know of to silence the alarm is to shutdown/reboot the server.

Edit: It has been brought to my attention that there are ways in which a disk can fail completely and not cause an audible alarm. This means it is possible that a disk can fail and you might never get a warning. So these generations of controllers should be avoided and never used with data that doesn't have thorough backups.


452x - These controllers work fine out of the box with any version after 8.3-RELEASE. I received the controller a little too late to get the changes into 8.3. Any hard drive that is not in a hardware RAID array is a "legacy disk". Everything works fine except SMART and serial numbers in the GUI. These issues are being addressed and hopefully will be supported soon. I will update this posting when I have more information. Edit: It appears that SMART is almost certainly never going to be supported as Highpoint has no way of querying disks on the controller and comparing them to actual devices. So these controllers should be avoided at all costs.

Overall, if you have the audible alarm enabled that should provide sufficient warning that a drive has failed. Since hot-swap is not supported you will need to shutdown the server, insert a new hard drive then power up the server and add the new hard drive to the zpool per the FreeNAS manual.

See ticket https://support.freenas.org/ticket/1932 for outstanding issues with some RocketRAID controllers.

2720SGL - I don't own one of these, but we've had many users that tried to use these. The sound great on the surface, but they aren't great once you have one. One of our users(JohnKnee) has provided a good discussion on various problems with these. In short, don't buy them...buy an IBM M1015 like we already recommend for everyone. If you already have it and can't return read his post here and understand that you are about to have potential serious problems.

------

If you have one of these controllers and are having issues or question I will attempt to provide assistance if I can.

Keywords: Highpoint RocketRAID 2210 2220 2224 2240 2300 2302 2304 2310 2312 2314 2314MS 2320 2322 2340 2522 3510 3520 3522 3530 3540 3560 4310 4311 4320 4321 4322 4520 4522
 
Last edited:

Joshua Parker Ruehlig

Hall of Famer
Joined
Dec 5, 2011
Messages
5,949
We might want to ask the FreeNAS developer to add an option to run smart tests on drives on a hpt controller. Not essential cause we can make it a cron job ourselves but it may be helpful to those who don't know how to run the command. I got it working with
Code:
smartctl -t short -d hpt,1/X/1 /dev/hpt27xx


This is on a recent daily build of FreeNAS 8.3.1, with my drive being /dev/daX
 

wobbel02

Cadet
Joined
Oct 3, 2012
Messages
1
Problems with RR3510

Very informative post. I found it while searching for a problem i have with my RR3510 controllers.
I have set up a machine with an Intel board Intel SE7230NH1, a Pentium D 925, 8GB of RAM and two volumes BACK and DATA. BACK is build from 3 drives with 2TB, 2 connected to the RR3510 as JBOD and 1 to the mainboard. DATA is build from 3 drives with 1TB, 2 connected to the RR3510 as JBOD and 1 to the mainboard. Currently i use FreeNAS 8.2.0-RC1-x64. The system is part of a windows domain. On each of the volumes there is a CIFS Share. Connection to the network is by 3 bundled Intel NICs.
Now i tried to copy some big files of about 1TB in size from another computer to the NAS. The performance was very bad and finally the transfer was aborted due to unrecoverable ZFS checksum errors. The errors were only shown for the drives connected to the RR3510. Behaviour was the same for both volumes.
I then changed the RR3510 to another one of same type. Behaviour was still the same, except that the transfer was not aborted and no unrecoverable errors were found.
I then removed the BBU from this RR3510 to let it use write-thru instead of write-back. The same tests showed no more checksum errors, the copy performance was more than doubled. But after a little more than 100GB of transferred data, the network connection broke down and the NAS shares were not reachable anymore from the source machine. Maybe another problem, not related to the NAS machine.
So my question: Is the controller cache of my RR3510s defective? Or is the controller caching mechanism in confict with the ZFS mechanisms? I don't believe in a hardware defect of the controller's cache, both have been running for quite a while in some windows servers, but it could be possible.
Can you give me any hints on this?
 

cyberjock

Inactive Account
Joined
Mar 25, 2012
Messages
19,526
Is the controller cache of my RR3510s defective?

Doubtful. The cache is ECC RAM. If it were to fail your controller firmware would likely spit out some kind of nasty error and not allow you to use it, or just not even initialize at all.

Or is the controller caching mechanism in confict with the ZFS mechanisms?

It's possible. My recommendation would be to disable the read and write cache. The read cache will hurt your performance, but a write cache MAY help your performance. Obviously, if you want to use the write back option of the write cache you should use your BBU.

Can you give me any hints on this?
Unfortunately, I don't have much advice to provide. Your issue likely isn't with the controller.

You may want to open a thread and see if anyone else has some ideas. You have a good amount of RAM, and the CPU should be good enough. That CPU is 6 years old. It is possible that the CPU chokes with the checksum calculations and that's why the server becomes unresponsive. Perhaps try running 2 drives in a mirror and see what happens when you copy large quantities of data. If it doesn't become unresponsive then the issue may be that the CPU can't handle the checksum calculations.
 

cyberjock

Inactive Account
Joined
Mar 25, 2012
Messages
19,526
No clue. I do not have one so I can't experiment with it. But considering that the website only provides drivers for Windows and Linux my guess is that FreeBSD is not supported at all :(
 

SnorreSelmer

Explorer
Joined
Sep 7, 2011
Messages
58
I'm running a FreeNAS server (8.2.0 b4) and I've struggled with various transfer performance-issues (SAMBA transfer-speeds are ~30MB/s and when I stream HD content off the server it occationally stutters badly for .5-1 sec before it gets well again).
Server-stats in the signature.

I've tried all kinds of things (SAMBA-settings and sysctls) and I still can't find the perp, so I'm wondering if I should install the official HighPoint driver for the controller to see if that helps. Where should I place the driver and how can I make sure it's active?

noobsauce80, do you have any experience with this? Any input on what to do or not do?
 

cyberjock

Inactive Account
Joined
Mar 25, 2012
Messages
19,526
What do you mean by "install the official highpoint driver"? The driver that FreeNAS uses IS the driver from Highpoint. If the driver wasn't loaded you wouldn't have access to the hard drives on the highpoint controller.

Things I'd check:

1. Are you using the 64 bit version of FreeNAS. If you aren't you are hurting performance because you aren't using all 8GB of RAM and prefetching will be disabled.
2. When streaming content that is stuttering what is CPU usage for FreeNAS doing?
3. 8GB of RAM is a tad small for that much storage space. The thumbrule is 1GB per TB of data. A friends' Q9400 couldn't go above 8GB of RAM though, so you may be maxed out.
4. You may have a hard drive failing. Have you run SMART tests on your hard drives lately?
5. Also does the content stutter at the exact same point every time it is played? This may be hard to test because FreeNAS has a cache and your media player probably does too.
6. What program are you using that stutters? If you use a different program, does the content still stutter?
7. What network card are you using? Intel cards are the performance king for FreeNAS. I've seen performance increases just by installing an Intel NIC.
8. Any reason you are using a beta version of FreeNAS over the full version? I'd recommend you upgrade to 8.3 unless there's something broken for you in the latest.
 

SnorreSelmer

Explorer
Joined
Sep 7, 2011
Messages
58
What do you mean by "install the official highpoint driver"? The driver that FreeNAS uses IS the driver from Highpoint. If the driver wasn't loaded you wouldn't have access to the hard drives on the highpoint controller.
There's a driver for the controller in FreeNAS yes, but I was wondering if the driver from HighPoint might give better performance (I don't have access to the server at this exact moment so I can't test it right away).

1. Are you using the 64 bit version of FreeNAS. If you aren't you are hurting performance because you aren't using all 8GB of RAM and prefetching will be disabled.
I run 64-bit FreeNAS.

2. When streaming content that is stuttering what is CPU usage for FreeNAS doing?
By the looks of it, next to nothing. If the server had encountered bad data that needed to be rebuilt then I expect I'd have seen it on the CPU load-meter.

3. 8GB of RAM is a tad small for that much storage space. The thumbrule is 1GB per TB of data. A friends' Q9400 couldn't go above 8GB of RAM though, so you may be maxed out.
It is? I know that the old (incorrect) rule of thumb for deduping was 1GB RAM per 1TB HDD, but I never heard that one about just general storage needing so much RAM. The motherboard (an XFX uATX board) is capped at 8GB so I can't put in more.

4. You may have a hard drive failing. Have you run SMART tests on your hard drives lately?
Haven't done that, but I scrub regularly, and I expect HDD issues to show up as problems when scrubbing. I could of course be wrong.

5. Also does the content stutter at the exact same point every time it is played? This may be hard to test because FreeNAS has a cache and your media player probably does too.
As far as I can tell, it's totally random. I can play the same movie several times and it will not stutter at the same spot twice, but it will always stutter (and yes, the movies are WAY bigger than what would fit in RAM, so the movie can't be all buffered in RAM). Also, the bit-rate of the movie doesn't seem to be a factor either.. My highest bit-rate movie (Minority Report @ 5MB/sec avg, 7MB/sec peak) stutters just as much/little as my lower bitrate movies (all the movies on this server are raw AVCHD streams w/ DTS-MA or Dolby Digital HD sound).

6. What program are you using that stutters? If you use a different program, does the content still stutter?
It's played on an AC Ryan media-streamer. The box runs some kind of Linux, has 256MB of streaming-buffer RAM and that's all I know.

7. What network card are you using? Intel cards are the performance king for FreeNAS. I've seen performance increases just by installing an Intel NIC.
The Realtek RTL8111C on-board NIC. If all else fails I'll look into getting an Intel NIC.

8. Any reason you are using a beta version of FreeNAS over the full version? I'd recommend you upgrade to 8.3 unless there's something broken for you in the latest.
I know. I was surprised when I checked the server the other day and saw that I hadn't even updated to 8.2.0 RELEASE.. :/ I'm upgrading the server to 8.3.0 RELEASE on Sunday (if all goes to plan).

Just as a closing statement, the reason why I suspect it to be related to the HighPoint controller is that the server ran just fine before I installed it. The RaidZ-configuration is the same, the RAM is the same, the CPU is the same. If it turns out to be that I have too little RAM then at least I've learned something new.
 

cyberjock

Inactive Account
Joined
Mar 25, 2012
Messages
19,526
There's a driver for the controller in FreeNAS yes, but I was wondering if the driver from HighPoint might give better performance (I don't have access to the server at this exact moment so I can't test it right away).

You won't be able to load the driver from the Highpoint website without disabling the kernel driver. AFAIK to do that you'll have to recompile FreeNAS' kernel.


By the looks of it, next to nothing. If the server had encountered bad data that needed to be rebuilt then I expect I'd have seen it on the CPU load-meter.

Not necessarily. If there are bad sectors the hard drive will stall. FreeNAS will not attempt to rebuild the data using parity/mirrors until the hard drive says it could not read the sector. Some people have had speeds of less than 1MB/sec with 1 dying hard drive. Streaming HD video isn't exactly high performance. 5MB/sec is somewhat typical. So your server clearly has performance issues if it is not able to stream a single video and nothing else.

It is? I know that the old (incorrect) rule of thumb for deduping was 1GB RAM per 1TB HDD, but I never heard that one about just general storage needing so much RAM. The motherboard (an XFX uATX board) is capped at 8GB so I can't put in more.

Actually, per the FreeNAS 8.3 release notes, dedup is 3-5GB per TB of data! The FreeNAS manual used to say 1GB of RAM per 1TB of data.

Haven't done that, but I scrub regularly, and I expect HDD issues to show up as problems when scrubbing. I could of course be wrong.

Scrubs should give you an indication of a failing hard drive, but SMART tests are better for finding failing drives. Sometimes scrubs will only take longer than they should. In your case, you might not even notice unless you see a sudden sharp increase in scrub time. I don't monitor my scrub times closely enough to tell the difference between 8h and 9h for a scrub.

As far as I can tell, it's totally random. I can play the same movie several times and it will not stutter at the same spot twice, but it will always stutter (and yes, the movies are WAY bigger than what would fit in RAM, so the movie can't be all buffered in RAM). Also, the bit-rate of the movie doesn't seem to be a factor either.. My highest bit-rate movie (Minority Report @ 5MB/sec avg, 7MB/sec peak) stutters just as much/little as my lower bitrate movies (all the movies on this server are raw AVCHD streams w/ DTS-MA or Dolby Digital HD sound).

Odd. But if you have a failing disk you can expect unpredictable results.

It's played on an AC Ryan media-streamer. The box runs some kind of Linux, has 256MB of streaming-buffer RAM and that's all I know.

I've never heard of or used it. It could be an issue with the device. Have you tried streaming video through a computer? If you can play your videos with something like VLC then then problem is more likely the box.

Edit: I just looked up AC Ryan, and they declared bankruptcy a few months ago. Sounds like they either are a crappy product or there was no market for their product. I really think you should try a different device such as a computer before you start looking at the server for the cause of the problems.

The Realtek RTL8111C on-board NIC. If all else fails I'll look into getting an Intel NIC.

Are you running gigabit everywhere? 100mb isn't an optimal solution for streaming HD content, but it should work.

Just as a closing statement, the reason why I suspect it to be related to the HighPoint controller is that the server ran just fine before I installed it. The RaidZ-configuration is the same, the RAM is the same, the CPU is the same. If it turns out to be that I have too little RAM then at least I've learned something new.

Umm.. so did you originally have 6 drives, then added 4 more later? You are aware that adding singular disks to a zpool can remove parity as described in my guide(link in the sig). Being that your sig says 14.5TiB of space that would mean a single vdev of RAIDZ2. If you added the 4 drives later you are running them in a RAID0 vdev. That means a failure of any of the 4 disks you added will result in a loss of all data in your zpool.
 

SnorreSelmer

Explorer
Joined
Sep 7, 2011
Messages
58
You won't be able to load the driver from the Highpoint website without disabling the kernel driver. AFAIK to do that you'll have to recompile FreeNAS' kernel.
I'd rather mental-floss with barb wire...

Not necessarily. If there are bad sectors the hard drive will stall. FreeNAS will not attempt to rebuild the data using parity/mirrors until the hard drive says it could not read the sector. Some people have had speeds of less than 1MB/sec with 1 dying hard drive. Streaming HD video isn't exactly high performance. 5MB/sec is somewhat typical. So your server clearly has performance issues if it is not able to stream a single video and nothing else.
That's my point too.. When I scrub the server, the drives operate steadily at 30MB/sec each, but that's half of what I would expect (even though the drives are of the Low Power type). And yes, I'm the only user of the server.

Actually, per the FreeNAS 8.3 release notes, dedup is 3-5GB per TB of data! The FreeNAS manual used to say 1GB of RAM per 1TB of data.
Need to look at that then.

Scrubs should give you an indication of a failing hard drive, but SMART tests are better for finding failing drives. Sometimes scrubs will only take longer than they should. In your case, you might not even notice unless you see a sudden sharp increase in scrub time. I don't monitor my scrub times closely enough to tell the difference between 8h and 9h for a scrub.
I'll look into that too.

I've never heard of or used it. It could be an issue with the device. Have you tried streaming video through a computer? If you can play your videos with something like VLC then then problem is more likely the box.
The box plays the HD movies perfectly from USB HDD or internal drive, just not over the network.. Could be a poor networking sub-system in the Linux, but I kinda doubt that (unless AC Ryan intentionally fiddled with knobs that should be left alone).


Edit: I just looked up AC Ryan, and they declared bankruptcy a few months ago. Sounds like they either are a crappy product or there was no market for their product. I really think you should try a different device such as a computer before you start looking at the server for the cause of the problems.
Will do. The box has huge potential, but AC Ryan failed to commit to software development and maintenance. Between the last official firmware and latest beta firmware are ~1500 builds, and the only listed fixes are "New NTFS driver" and "Fixed a YouTube problem"... *sigh*

Are you running gigabit everywhere? 100mb isn't an optimal solution for streaming HD content, but it should work.
Yep. There's only one device between the server and the media-streamer, and that's a Cisco router with Gigabit ports. Everything is of course wired with Cat5e or Cat6 cables.

Umm.. so did you originally have 6 drives, then added 4 more later? You are aware that adding singular disks to a zpool can remove parity as described in my guide(link in the sig). Being that your sig says 14.5TiB of space that would mean a single vdev of RAIDZ2. If you added the 4 drives later you are running them in a RAID0 vdev. That means a failure of any of the 4 disks you added will result in a loss of all data in your zpool.
I tore down the array and built a new one from scratch. It used to be a 4+2 array, now it's 8+2, not a (4+2)+(2+2). From what I understand from reading the ZFS docs is that you can't add vdevs of different raidz-levels together in the same pool, so I built a completely new one.
 

cyberjock

Inactive Account
Joined
Mar 25, 2012
Messages
19,526
Will do. The box has huge potential, but AC Ryan failed to commit to software development and maintenance. Between the last official firmware and latest beta firmware are ~1500 builds, and the only listed fixes are "New NTFS driver" and "Fixed a YouTube problem"... *sigh*

FreeNAS has similar issues. Their changelog only lists the big ticket items. There's no way they could list everything they did, otherwise the changelog would be as big as the manual. That's one of the reasons why when someone has an odd obscure problem and are using a version prior to 8.2 I often ask why they haven't upgraded. So much stuff is optimized and fixed that I can't believe how many people come posting in the forum without trying the simple stuff first. Upgrading really is so easy.

I tore down the array and built a new one from scratch. It used to be a 4+2 array, now it's 8+2, not a (4+2)+(2+2). From what I understand from reading the ZFS docs is that you can't add vdevs of different raidz-levels together in the same pool, so I built a completely new one.
Well, you CAN mix different raidz levels, but you have to know what you are doing and understand the risks. For the most part it's not recommended because the added complexity adds ALOT of room for errors.
 

ramius

Dabbler
Joined
Oct 30, 2012
Messages
17
22xx & 23xx - These controllers do NOT mount hard drives that are not part of a RAID array as a "Legacy Disk". To be mounted as a "Legacy Disk" they MUST contain a valid partition table with at least one partition. Unfortunately ZFS partitions are NOT identified as valid partitions.

The workaround is to create a partition on the hard drive that is FAT16, FAT32 or NTFS of any size. In my case I created a 100MB partition that is FAT16. If at any time your FAT16 partition table entry is deleted or otherwise corrupted you WILL lose the ability to access your ZFS partition when connected to a Highpoint controller.

Maybe it's my controller, or the bios version, but I've a RR2300 controller. This morning i booted the system with a drive ataches to the controller, the drive has never been conected to any RR controller. The drive had a NTFS partition and, as you said, the drive was mark by the controler as a legacy drive. I added the drive to freenas and after transfering the information, I wiped the drive and formated the drive with the freenas GUI as a zfs single drive. After rebooting the system again, the drive still apears in the controler menu as a legacy drive and in freenas as "HPT DISK 0_0" drive. My concern is, will this drive dissaper or will it stay as a lagacy drive?
 

cyberjock

Inactive Account
Joined
Mar 25, 2012
Messages
19,526
I'm not sure what you define as "wiped". If you wiped the free space, then the NTFS partition is still there. Can you post your partition table for one of your drives? My guess is you still have an NTFS or FAT partition on the drive.

Edit: As an aside, and I plan to upgrade the guide soon, Highpoint controllers are a bad way to do business right now. One of the developers and I looked at adding SMART support as well as serial numbers being provided in the GUI. Neither one will implement easily. Right now you have NO way to run FreeNAS with SMART support without some reconfiguring because the Highpoint CLI doesn't work out-of-the-box.
 

ramius

Dabbler
Joined
Oct 30, 2012
Messages
17
I used the option from the GUI to detach the volume and mark the disk as new. After that I used the Volume Manager to add a new drive with a ZFS filesystem.

Here is the partition table from the drive

Code:
[root@freenas] ~#  fdisk da3
******* Working on device /dev/da3 *******
parameters extracted from in-core disklabel are:
cylinders=30401 heads=255 sectors/track=63 (16065 blks/cyl)

Figures below won't work with BIOS for partitions not in cyl 1
parameters to be used for BIOS calculations are:
cylinders=30401 heads=255 sectors/track=63 (16065 blks/cyl)

Media sector size is 512
Warning: BIOS sector numbering starts with sector 1
Information from DOS bootblock is:
The data for partition 1 is:
sysid 238 (0xee),(EFI GPT)
    start 1, size 488397167 (238475 Meg), flag 80 (active)
        beg: cyl 0/ head 1/ sector 1;
        end: cyl 1023/ head 255/ sector 63
The data for partition 2 is:
<UNUSED>
The data for partition 3 is:
<UNUSED>
The data for partition 4 is:
<UNUSED>
 

SnorreSelmer

Explorer
Joined
Sep 7, 2011
Messages
58
Ok, time for a confession. I just finished watching 30 minutes of Minority Report (my highest bit rate movie) after upgrading to 8.3.0 RELEASE and there was not a single snag. Not one! Apart from the upgrade I turned on Autotune, but it seems to be working great now. Apologies for polluting your thread with what turned out to be controller unrelated issues. The thread has been very interesting none the less.

Ps: I actually did the upgrade from my Galaxy S3 and while a bit painful, it actually worked.
 

ShaharHD

Dabbler
Joined
Nov 2, 2012
Messages
20
We might want to ask the FreeNAS developer to add an option to run smart tests on drives on a hpt controller. Not essential cause we can make it a cron job ourselves but it may be helpful to those who don't know how to run the command. I got it working with
Code:
smartctl -t short -d hpt,1/X/1 /dev/hpt27xx

This is on a recent daily build of FreeNAS 8.3.1, with my drive being /dev/daX

With this I was actually finally able to read smart data of my drives connected to the RR 2720SGL card!
Code:
smartctl -i -d hpt,1/X/1 /dev/hpt27xx

and for future reference ... now I have this: http://sourceforge.net/apps/trac/smartmontools/wiki/Supported_RAID-Controllers

to continue with your idea, is there a way to correlate between the /dev/daX devices to the 1/X/1 mapping?
as in my case I have the following:
Code:
/dev/da0     /dev/da1p1   /dev/da2     /dev/da2p2   /dev/da3p1   /dev/da4     /dev/da4p2   /dev/da5s1   /dev/da5s2   /dev/da5s3
/dev/da1     /dev/da1p2   /dev/da2p1   /dev/da3     /dev/da3p2   /dev/da4p1   /dev/da5     /dev/da5s1a  /dev/da5s2a  /dev/da5s4

but the "X" I'm getting to work are: 2,3,4 & 7,8 (5 drives are connected out of the 8) - 3 to the one SAS port and 2 to the other (other drive bays are empty at the moment).

if we can do the correlation, then there's a way to modify the smartctl command for those devices to allow also for the Web GUI to have the correct information.
 

cyberjock

Inactive Account
Joined
Mar 25, 2012
Messages
19,526
to continue with your idea, is there a way to correlate between the /dev/daX devices to the 1/X/1 mapping?
as in my case I have the following:
Code:
/dev/da0     /dev/da1p1   /dev/da2     /dev/da2p2   /dev/da3p1   /dev/da4     /dev/da4p2   /dev/da5s1   /dev/da5s2   /dev/da5s3
/dev/da1     /dev/da1p2   /dev/da2p1   /dev/da3     /dev/da3p2   /dev/da4p1   /dev/da5     /dev/da5s1a  /dev/da5s2a  /dev/da5s4

but the "X" I'm getting to work are: 2,3,4 & 7,8 (5 drives are connected out of the 8) - 3 to the one SAS port and 2 to the other (other drive bays are empty at the moment).

if we can do the correlation, then there's a way to modify the smartctl command for those devices to allow also for the Web GUI to have the correct information.

As far as I know there is no way to correlate physical ports to /dev/daX devices at this time. The CLI doesn't even appear to provide that information. As I said in ticket 1932 it may be impossible to ever get SMART features to work correctly from within the GUI since there is a basic need to correlate the daX to the physical port so that the GUI can give you the correct serial numbers in the GUI as well as setup the SMART settings the way the user specifies in the GUI. As it currently stands if the CLI were installed I'd be forced to create my own config file manually for smartctl and add it to the startup of the OS and have to accept the fact that I will have to manually control it myself between FreeNAS upgrades.

I put in a ticket with Highpoint asking for assistance in identifying the daX to the physical port, but they haven't responded at all(my ticket was submitted Nov 21st). I'd at least like an answer from them regarding if it is possible somehow or not. If not then perhaps I could coax them into making that feature a part of the CLI.

It's a real bummer too because I know 3 different people that have bought Highpoint controllers and none of them want to switch to FreeNAS as long as this issue exists. This is the only issue holding them back from migrating to FreeNAS right now.
 

cyberjock

Inactive Account
Joined
Mar 25, 2012
Messages
19,526

TnT

Cadet
Joined
Mar 25, 2013
Messages
4
I hope I'm allowed to Necro old posts, as this one is fantastic! I am currently using a HighPoint RocketRaid 2314 and this post explained many of the issues I am experiencing.

Ultimately, I am looking to use my NAS for a few things
- NFS shares presented to my ESXi box for a few basic VM's.
- Video Storage (For streaming video)
- Document storage

I have noticed one of the channels on my RR 2314 is being soft reset. What could be the cause of this? (Bad Hard Drive? Bad Driver? Bad Firmware?). Before I start loading on a lot of important data, I want to ensure my NAS is as stable as possible - this soft reset is causing me concerns.
What can I start looking for to try to narrow it down?

Do you have any experience with a HighPoint RocketRaid 644L? I am finding these online for ~$100-$120 and they might be a good replacement for my RR 2314. Edit: On further investigation it looks like these cards are not supported on FreeBSD. What card would you recommend?
 
Status
Not open for further replies.
Top