ARECA Owner's Thread (SAS/SATA RAID Cards)

Discussion in 'SSDs & Data Storage' started by odditory, Jan 5, 2010.

  1. Ondrej

    Ondrej n00b

    Messages:
    19
    Joined:
    Jun 15, 2017
    Hello to everybody,
    so did it. I backed up everything and I had to format the disk array to a 64KB Cluster size.

    And then I tried to initialize the new disk and after a week of migration I extended the partition with another full 10TB (9.3TB) and everything is fine :)

    Now I run on Raid6 and 7 units of 10 TB (home storage)
     
    arnemetis likes this.
  2. arnemetis

    arnemetis 2[H]4U

    Messages:
    2,707
    Joined:
    Aug 2, 2004
    So I have my arrays set up to do a monthly scrub. I had an unrelated issue today and decided to check the log anyway to make sure something odd wasn't going on. The latest scrub of my 8x 8tb array found 2 errors. What can this mean? I can't find any additional information, all previous checks pulled up 0 errors. Did the scrubbing fix the errors?
    upload_2019-7-2_17-58-56.png
    HD Sentinel shows no issues with the drives. I hope I'm just being paranoid here!
     
  3. Snowknight26

    Snowknight26 [H]ardness Supreme

    Messages:
    4,163
    Joined:
    May 8, 2005
    This is a long shot but what stick of RAM did you use that allowed your ARC-1882IX-24 to use 8GB (more than 4GB advertised maximum)? I tried a Micron MT18KSF1G72PZ-1G6E1 (which is 1Rx4) without success.
     
  4. Ondrej

    Ondrej n00b

    Messages:
    19
    Joined:
    Jun 15, 2017
    Hi Snowknight26 i have a
    Transcend - TS1GLK72V3H
    https://www.ebay.com/sch/i.html?_from=R40&_trksid=m570.l1313&_nkw=TS1GLK72V3H&_sacat=0 and work fine :)

    I tried a lot of brands and different and without success. Here this type of advice has been given to every kid on ebay when I was buying a 2GB module.
     
  5. Snowknight26

    Snowknight26 [H]ardness Supreme

    Messages:
    4,163
    Joined:
    May 8, 2005
    Dual rank, interesting. That goes against what the manual says:

    But I'll give a 2Rx8 module a shot.

    Edit:
    After posing my original question I found an Areca FAQ article that confirmed 8GB support.

    http://faq.areca.com.tw/index.php?action=artikel&cat=14&id=359&artlang=en&highlight=8gb
     
  6. Snowknight26

    Snowknight26 [H]ardness Supreme

    Messages:
    4,163
    Joined:
    May 8, 2005
    For those that are curious, I was able to get 8GB in my ARC-1882ix-24 using a stick of ADATA ADDE1333W8G9-BMIE, which matches what Areca's FAQ says: 2Rx8 unbuffered.

     
    Lifespeed likes this.
  7. WoooT

    WoooT n00b

    Messages:
    43
    Joined:
    Aug 1, 2010
    Iam running 4 disk Raid 6 on Areca 1220, apparently i have degraded raid set and one disk is faulty according to online raid manager.
    i bought a new disk to replace the faulty one but now that i have done so i put the faulty one in my desktop to see the issue, it looks like the disk is all fine, Smart values are OK when i've done fast check
    and i will continue do more long test to be sure that the disk is OK.

    Dide any of you had same kind of problem with areca raid set being degraded and marking healthy disk as bad?
    what could be the cause of it?
    ive hade couple of power outs previously where the server shut off could that be the cause? improper shutdown-.
     
  8. Blue Fox

    Blue Fox [H]ardForum Junkie

    Messages:
    11,693
    Joined:
    Jun 9, 2004
    What do the logs show? If the disk was timing out, then the card will drop the drive eventually.
     
    arnemetis likes this.
  9. WoooT

    WoooT n00b

    Messages:
    43
    Joined:
    Aug 1, 2010
    This

    2019-08-28 16:26:49 ARC-1220-VOL#00 Start Rebuilding
    2019-08-28 16:26:46 Raid Set # 00 Rebuild RaidSet
    2019-08-28 16:24:44 Proxy Or Inband HTTP Log In
    2019-08-28 16:11:59 H/W Monitor Raid Powered On
    2019-08-26 23:23:13 Proxy Or Inband HTTP Log In
    2019-08-20 01:07:08 H/W Monitor Raid Powered On
    2019-08-13 02:57:34 IDE Channel 4 Device Failed
    2019-08-13 02:57:34 Raid Set # 00 RaidSet Degraded
    2019-08-13 02:57:34 ARC-1220-VOL#00 Volume Degraded
    2019-07-19 19:31:14 H/W Monitor Raid Powered On
     
  10. mwroobel

    mwroobel [H]ardness Supreme

    Messages:
    4,948
    Joined:
    Jul 24, 2008
    What brand and model are the drive(s) in the array? How long has the array been in use, and is this the first time it has happened? What were the SMART readings from the drive that dropped out of the array? What firmware version if your card currently flashed to? If this is the first time and the array has been up for a while, the drive is most likely on its way out (Failed vs Timed-Out) or you may have a power problem. If you have a replacement disk use that, and then you have the time to run more stressful tests on the drive that failed. Improper shutdowns can cause problems, but usually drives not spinning up or not having a battery backup and the card not having time to flush the cache to disk. You should consider a UPS for any RAID array so it can shutdown gracefully in case of power issues, which it seems you are having.
     
  11. WoooT

    WoooT n00b

    Messages:
    43
    Joined:
    Aug 1, 2010

    I have 4 of these 4tb drives, HGST HDN724040ALE640 they have been running for 3 years almost they have all been purchased at the same time.

    Smart values from ARECA on these disks are quite similar

    SMART Read Error Rate 100(16)
    SMART Spinup Time 121(24)
    SMART Reallocation Count 100(5)
    SMART Seek Error Rate 100(67)
    SMART Spinup Retries 100(60)

    SMART Read Error Rate 100(16)
    SMART Spinup Time 127(24)
    SMART Reallocation Count 100(5)
    SMART Seek Error Rate 100(67)
    SMART Spinup Retries 100(60)

    SMART Read Error Rate 100(16)
    SMART Spinup Time 127(24)
    SMART Reallocation Count 100(5)
    SMART Seek Error Rate 100(67)
    SMART Spinup Retries 100(60)

    Controller Name ARC-1220
    Firmware Version V1.49 2010-12-02
    BOOT ROM Version V1.49 2010-12-02
    Serial Number Y805CAAHAR200552
    Unit Serial #
    Main Processor 500MHz IOP333
    CPU ICache Size 32KBytes
    CPU DCache Size 32KBytes/Write Back
    System Memory 256MB/333MHz/ECC

    The "bad" disk had the same values, unfortunately i didn't save it before taking it out.

    iam doing a longtest now on the "bad" drive,


    Edit:

    i actually saved the SMART for the BAD disk i can now see the Spinuptime is very high.
    Namnlös.png

    Namnlös2.png

    The reason of the high temperature is, i have the disk on my desktop connected "externally" for testing.


    Edit #2
    I have done long test of the disk now, i used DFT, Drive fit test by HGST and both short and longtest is OK.

    DO i dare to put the disk back in the raidset?
     
    Last edited: Aug 29, 2019
  12. Lifespeed

    Lifespeed n00b

    Messages:
    62
    Joined:
    May 12, 2013
    Thanks, just got a stick to try out as part of my 1882ix-16 RAID6 drive upgrade.

    Edit: I got a used 8GB stick of Ebay from a decomissioned server for $16. Exact ADATA part number you mention.

    My media server boots much faster now, and file transfers between the RAID array and a local SSD are blazing fast - 600MB/S+ on a nearly-full array. Note that RAM insertion is in an awkward position. Double check for good seating and latching, it won't work otherwise as I found out.
     
    Last edited: Sep 15, 2019
  13. Lifespeed

    Lifespeed n00b

    Messages:
    62
    Joined:
    May 12, 2013
    It's time to finally make a move on a hard drive upgrade to my 1882ix-16. Yeah, it is no longer the newest card, but if I can find compatible drives I don't see a reason to upgrade the card, especially given the cost.

    I want to use 12TB Hitachi with 4K native sector drive base on excellent results from my own experience, and Backblaze had excellent results as well. The Toshiba 14 and even 16TB are tempting, but reliability is huge in this application and HGST still seems to be the winner there.

    So I'm going to update the firmware on my 1882-ix16 card, plug in the fleabay 8GB ram stick (thanks Snowknight26 ), and build a new array to retire the legendary Hitachi 3TB drives. They've served me well since 2010, time for all 12 of them to be replaced. While I still have 4 port left on the array to migrate too ;)

    Of course I have no idea on TLER compatibility, but some models are considered enterprise or NAS. Any suggestions on array testing procedures and software, stripe sizes, etc, are welcome. I'm going to try and check with the drive retailer about return policy, as I'll have to get 4 drives minimum to start with to transfer my 30TB data from the old array. Even then I'll have to migrate RAID5 to RAID6, and then add a drive to make this a true storage upgrade.
     
  14. Lifespeed

    Lifespeed n00b

    Messages:
    62
    Joined:
    May 12, 2013
    Here's a link to the large, modern helium drives with an ARC-1882ix-16 thread. I'm going to buy a few WD DC HC530 14TB drives, which apparently are Hitachi Helioseal thanks to the merger. If you have any advice, please join the conversation. It may be relevant to quite a few of we Areca owners.
     
  15. pgjensen

    pgjensen Limp Gawd

    Messages:
    323
    Joined:
    Apr 20, 2006
    Can someone recommend a cheap 8GB ECC DDR3-1866 stick for the Areca 1883i card?

    *Edit: looks like they aren't any 1866 sticks compatible with this card, so opted for a 1600 stick at $23 shipped on ebay.
     
    Last edited: Sep 18, 2019
  16. mwroobel

    mwroobel [H]ardness Supreme

    Messages:
    4,948
    Joined:
    Jul 24, 2008
    My suggestion would be to by a 12TB SAS drive like the HUH721212AL5200, which is on the HCL for the 1882.
     
  17. Lifespeed

    Lifespeed n00b

    Messages:
    62
    Joined:
    May 12, 2013
    Already pulled the trigger on 14TB Western Digital DC HC530 "Hitachi" SATA drives. Couldn't let that extra storage escape me during my once-a-decade RAID hard drive migration. These are data center drives with rotational vibration sensors, motors on both ends of the spindles, and probably some other heavy-duty features. I'm only missing some speed without SAS, but these are going to be ridiculously fast anyway. Areca support thought they would be compatible, I'll report back soon.
     
  18. Lifespeed

    Lifespeed n00b

    Messages:
    62
    Joined:
    May 12, 2013
    The 4 X 14TB Western Digital drives in my last 4 slots of ARC-1882ix-16 are building a RAID5 array currently at 17% after 12 hours. 42GB usable storage compared to my 12 X 3TB Hitachi RAID6 at 30GB. This could take a few days. After testing and moving my data I'll pull the old 3TB drives, buy a couple more 14TB drives, migrate to RAID6, then expand with one more drive to 56TB. Should last me a while. A home media server can be an expensive hobby, but sure is nice to have!

    Any suggestions on software to stress-test the new array?

    RAID_init_14TB.PNG
     
    Last edited: Sep 19, 2019
  19. maxxmusterr

    maxxmusterr n00b

    Messages:
    12
    Joined:
    Jan 31, 2018
    Looking forward to your findings.
    And thanks for the RAM tipp. Just make sure you have a healthy battery connected to your card :) But I am sure you already know. 16USD - Good deal :D ....

    It's not exactly a stress test per se. But there is a tool called h2testw.
    https://www.heise.de/download/product/h2testw-50539
    What is does is basically: it makes small junks/pieces and write it to the disk (until it's full). Then you can let it verify every piece again you have written, if it is really okay.
    You can let it do that in an endless loop and it does it itself in the background. If you let that do it's job for a week or two, then I still would call that a pretty neat stress test. At least you have written, read and verified a lot of data.
    The tool was designed for testing (fake) USB sticks, before use (because there are some where the capacity is faked/blew up) - another use case. But it does it job also with HDs.
     
  20. Lifespeed

    Lifespeed n00b

    Messages:
    62
    Joined:
    May 12, 2013
    I can't read German, not sure that tool will do me much good. I have never had a backup battery on this RAID card. I used to have a UPS, but grew tired of failed batteries (which I could deal with) and failed electronics. I will get a new UPS for the server with the RAID array eventually. I suppose I should get a RAID battery, but that would just be another maintenance item.

    The 56TB (42TB usable) RAID5 array has finished. I decided to postpone stress testing, and instead will use the transfer of 30TB of data from my old array, RAID level migration to RAID6, and array expansion with another drive as an opportunity for the hard drives to reveal their compliance with TLER/ERC. As time allows during this process I'll run some write/readback tests as well. I have a backup of critical data, and non-critical media is still on the old array so not much risk here.
     
  21. maxxmusterr

    maxxmusterr n00b

    Messages:
    12
    Joined:
    Jan 31, 2018
    Lifespeed i's English too if you open it (toggle key)

    But in general if you go to an individual drive - Do you see TLER/ERC//CCTL as enabled and how many seconds does it display?

    Battery i still would suggest if the system freezes (and you have to pull the plug) so it still loses power off course. I had it a several times over the years, so to be super sure i did buy one.
    I am very cautious, because I had once a silent filesystem failure in the background, which left me a few unreadable files, which the system did not detect.
     
  22. Lifespeed

    Lifespeed n00b

    Messages:
    62
    Joined:
    May 12, 2013
    maxxmusterr, if by "go to an individual drive" you mean the web interface to the 1882ix-16, I don't see TLER/ERC/CCTL terminology shown. Error Recovery Control is shown as disabled for the new 14TB disk array. However, Error Recovery Control is also shown as disabled for the Hitachi 3TB array that I have been running for 10 years. I thought the Hitachi 3TB were known to support this feature. In 10 years of use I have not had a disk drop out.

    To quote Areca support "ERC is a quite common feature on enterprise drives, and it is rare to get compatible issue on using enterprise drives." These drives are clearly sold as "enterprise".

    So I'm really not sure what to make of this, but I'll ask Areca and include screenshots.

    RAID_ERC_14TB.PNG
     
  23. arnemetis

    arnemetis 2[H]4U

    Messages:
    2,707
    Joined:
    Aug 2, 2004
    I don't have any particular experience with the issue you're troubleshooting, but I do have two arrays on an 1883 that I thought I'd show you what I have. Perhaps it will help? My hgst coolspin drives don't have ERC available (that array has been kicking since 2015) but my wd reds shucked from best buy easystore sales do have it enabled.
    upload_2019-9-24_13-47-12.png upload_2019-9-24_13-47-57.png
     
  24. maxxmusterr

    maxxmusterr n00b

    Messages:
    12
    Joined:
    Jan 31, 2018
    Yes - That's exactly the question in this thread.
    If it sees TLER/ERC/CCTL or not - if not displayed there I am thinking it's not working.... why not displaying it, when it works? Only explenation can be RAID FW or drive FW or a combination of them.
    And when i tested the arrays with heavy copying and h2test2 then yeah - the controller dropped the 12TB harddrives at some point, where the 4 and 6TB arrays were rock solid over years (and still would be) in the same machine.

    On my old 4TB and 6TB Seagate NAS drive arrays it was always 7.0/7.0 seconds, which gave me the sign it's okay and working.
     
  25. Lifespeed

    Lifespeed n00b

    Messages:
    62
    Joined:
    May 12, 2013
    The old array has been copied to the new array using Robocopy in Windows Powershell, a very convenient tool that preserves permissions and timestamps. Currently h2testw utility linked above is being run in the free space of the array to check for disk dropouts due to apparent incompatibility between WD DC HC530 and ARC-1882ix-16 firmware regarding the ERC command interface. The performance of the new disks with the old 1882ix-16 (upgraded to 8GB RAM) is impressive!

    I have sent Areca support the product manual for the DC HC530 hard drives, referencing pages 92, 96, 101 and 104 describing the command query interface for Error Recovery Control, which is clearly supported by these hard drives. Hopefully Areca finds it compelling to revise their firmware to work with these drives likely to be connected to their RAID cards. It is clear to me that while the drives support ERC, Areca firmware needs some updating to match.

    That said, with my use of the RAID5 (soon to be RAID6) I'm not finding the dropped-from-the-array type of incompatibility that was seen with the old WD consumer drives from a few years ago. I will also note my old HGST 3TB drives did not report ERC compatibility, yet ran for a decade without a drive ever dropping from the array.

    Soon I'll buy two more drives to expand from 4 X 14TB RAID5 to 6 X 14TB RAID6, probably 4 days for each of the two migrations. After which I'll decommission the old 12 X 3TB array and will hopefully be done with my storage fetish for a number of years.
     
    Last edited: Sep 29, 2019
  26. Lifespeed

    Lifespeed n00b

    Messages:
    62
    Joined:
    May 12, 2013
    I have been in communication with Areca regarding the inability of the ARC-1882i-16 (and probably their newer 1883 and 1884 models) to recognize and (presumably) configure ERC/TLER/CCTL with the Western Digital DC HC530 14TB drives. Areca support seems to want me to prove that these drives actually support ERC. Apparently sending them the product manual with references to the page numbers showing the implementation of ERC in drive firmware was not enough.

    So, what software utility can I use to query ERC status? Do I need to connect the drive(s) to a motherboard SATA port to query the ERC feature, or will it work through the Areca card? I get the impression Areca might address the compatibility issue on their end if they are shown irrefutable evidence the problem is theirs to fix. Certainly these drives are obvious candidates for RAID arrays, and Areca should support them. I'm just not sure how to go about providing this evidence.

    I think following through with this could be beneficial to the Areca user community in general, thanks for your help.
     
  27. mwroobel

    mwroobel [H]ardness Supreme

    Messages:
    4,948
    Joined:
    Jul 24, 2008
    On disks that fully implement the ATA-8 standard, the smartctl utility can be used to control the ERC behavior of many drives by setting the SCT Error Recovery Control (scterc) parameter:

    • Reading current settings:
      smartctl -l scterc /dev/sda
      SCT Error Recovery Control:
      Read: Disabled
      Write: Disabled

    • Changing the setting:
      smartctl -l scterc,150,150 /dev/sda
      SCT Error Recovery Control:
      Read: 150 (15.0 seconds)
      Write: 150 (15.0 seconds)
    ERC control needs to be set on the boot time and if hot-replacement been made.

    /sda above is used as a sample, you must use the address of your drive.
     
  28. Lifespeed

    Lifespeed n00b

    Messages:
    62
    Joined:
    May 12, 2013
    I did install a RAID backup battery per your suggestion. I may not go with a full UPS for the whole server and network for quite a while. I have had UPS' in the past, and have found them to be expensive, high-maintenance and failure prone. Electric power is reliable at my location so I find the cost/benefit ratio too high. Maybe after lithium ion UPS become more mainstream.
     
  29. Operations

    Operations n00b

    Messages:
    30
    Joined:
    Mar 11, 2019
    I am trying to replace the fan on my 1882-8i. But i have also removed the black metal. Could i use Thermal Grizzly paste or do i really need a thermal pad because of the die being "naked" and not able to handle the pressure without thermal pad?

    By "naked" i mean the LSI chip doesn't have a heatspreader like cpu's.
     
  30. pgjensen

    pgjensen Limp Gawd

    Messages:
    323
    Joined:
    Apr 20, 2006
    I tried an 8gb DDR3 ECC 2Rx8 PC3-12800 (1600) ram stick without success in my 1883ix-12 card. It said "SDRAM speed not enough." The RAM model was MT18KDF1G72PDZ-1G6E1. Anyone know of an 8GB stick for the 1883ix series that works?
     
  31. mwroobel

    mwroobel [H]ardness Supreme

    Messages:
    4,948
    Joined:
    Jul 24, 2008
    The DDR3 on my 1883's are PC3-14900 (1866). Your stick is not fast enough.
     
  32. pgjensen

    pgjensen Limp Gawd

    Messages:
    323
    Joined:
    Apr 20, 2006
    Do you have an 8GB stick? If so, what's the model number? Per the manual, it says:

    "
    One 240-pin DIMM socket for 2GB(default) up to 8GB DDR3-1866,1Rx8, ECC module (which doesn't seem to exist as 8gb 1Rx8)
    or 8GB, DDR3-1600, 2Rx8, ECC module (which is what I got)
    "
     
  33. mwroobel

    mwroobel [H]ardness Supreme

    Messages:
    4,948
    Joined:
    Jul 24, 2008
    The areca sheet says it supports a 240-pin DIMM socket for 2GB(default)
    - up to 4GB or 8GB DDR3-1866, 1RX8, or
    - up to 4GB or 8GB DDR3-1600, 2RX8, Unbuffered/ Registered ECC module

    ‘That being said, every one I have (2 and 4GB) only have 1866. It should work with a 1600 but you need to find out which specific stick Areca supports. Welcome to enterprise level memory support, especially when the manufacturer sells an overpriced OEM module.
     
  34. Operations

    Operations n00b

    Messages:
    30
    Joined:
    Mar 11, 2019
    Maybe better luck here :)
    It's maybe a bit of a silly question, but....

    I have upgrade the firmware on my Areca 1882 card to 1.56. Files located here:

    https://www.areca.us/support/download/adapter/fw/ARC-1882.zip

    Now as we all know, that gives you 4 files. I have selected them all in the menu Upgrade Firmware @ IP Address:82.

    Firmware is showing 1.56 after reboot. But when i go into the dos menu (by pressing TAB) the upperleft corner is still showing BIOS 1.24. Shouldn't this also show 1.56 now? I am not sure what it was saying before the upgrade. Should i try to flash the one file named BIOS through the DOS method?

    Example (from internet):

    http://blog.iamfilmguy.com/wp-content/uploads/2013/06/IMG_5117.jpg

    Here the BIOS says 1.22f, so again it is not my Areca card.
     
    Last edited: Nov 4, 2019
  35. Lifespeed

    Lifespeed n00b

    Messages:
    62
    Joined:
    May 12, 2013
    I think you update each file individually, not select them all in the browse for file window in your web browser. On my 1882 it shows current versions of all firmware including the BIOS.
     
  36. Operations

    Operations n00b

    Messages:
    30
    Joined:
    Mar 11, 2019
    As for as i know you cannot select them all in once. So i did select them individually :)
    So if you go in to the DOS menu, the left top corner shows your current version ? (probably 1.5*)
     
  37. Lifespeed

    Lifespeed n00b

    Messages:
    62
    Joined:
    May 12, 2013
    I haven't looked at the ARECA BIOS user interface, if that is what you mean by "DOS". I don't believe MS-DOS is involved with this at all. You have to update each file, not sure why you're having trouble.
     
  38. Operations

    Operations n00b

    Messages:
    30
    Joined:
    Mar 11, 2019
    Have you looked at the URL i posted? What i mean is when you reboot your computer and after POST you can press TAB/F6 to go into Areca menu. That is the menu i am talking about. In this menu the top left corner shows BIOS version. Like i said 1.24 in this case.

    Looking at the web interface only shows you the firmware and bootrom version when looking at system information. Both are showing 1.56.
     
  39. pgjensen

    pgjensen Limp Gawd

    Messages:
    323
    Joined:
    Apr 20, 2006
    Mine shows 1.24 bios. Updated 4x, once for each file. No need to reboot until all 4 are uploaded.
     
  40. Operations

    Operations n00b

    Messages:
    30
    Joined:
    Mar 11, 2019
    If yours is also showing 1.24 then that is probably normal. If feels a bit weird of your use an update file called BIOS, that the bios version number is still the same. But okee :)