New server with rocketraid 3560 and 24 hitachi 7k2000 on server 2003x64. Appears to be awesome with 3 raid 5s. Then every other day the server locks up with event vwr errors about sending resets to controller. On latest firmware and tried both drivers on highpoints site On the road so I can post more info if needed. I begged to build this server for my work now I feel like I am in hell. Highpoint has not responded to me yet. Any thoughts? Should I buy an aereco and return this?
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
RocketRaid 3560 nightmare 7k2000
- Thread starter Namu
- Start date
I would at minimum try giving that card extra cooling. There is a reason Highpoint cards are generally less expensive... because they do not have the same reputation as LSI, Adaptec, and Areca. A good indicator is the fact that they are not often OEM'd by the big storage players.
I tried lowering the temp of the server room to 62 degrees and it still did it. Here are the full details.
BIOS/Firmware Version: 1.2.15.23
Driver version: 1.2.25.9 (storport)
WebGUI version: 1.0-080120
Operating System: Windows Server 2003 x64 Standard
Motherboard Model: Asus P8P67 PRO
System BIOS version: 0701
System memory size: 8GB
HDD Model: Hitachi HDS722020ALA330
HDD Firmware version: _______
RAID Configuration: 3 Raid 5s with 8 drives in each
Case: NORCO RPC-4224 4U
Motherboard: ASUS P8P67 PRO
PSU: CORSAIR CMPSU-850TX 850W
Raid: RocketRaid 3560
Other configuration information:
When using the VSS driver (1.2.28.28) Raid drives would disconnect from windows and EVENT was logged:
Source: hptiop Event ID: 9
The device, \Device\Scsi\hptop1, did not respond within the timeout period
This Message will repeat about 2 times and then once or twice later on with 20mins in b/w. I can still login to windows and kind of use it(i can access the C: drive and pull up event manager but anything else will freeze it including trying to shutdown.
When using 1.2.25.9 driver, entire server locks up and events are made every 30Seconds until forced reboot:
source: hptiop Event:129
Reset to device, \Device\RaidPort0, was issued.
With this driver I can do nothing at all on the server, complete freeze the second it happens.
I can still connect to the out-of-band managagemnt webpage for the raid card and no events are logged on it at all. If I submit a rescan on the drives the webpage will then lock up and I can no longer get to it.
I have 2 RocketRaid2320 and they never had issues with these drives.
I have tried switching PCI slots. swaping those 2 drivers, spreading the drives power across the PSUs molex and Sata connectors. Adding the PCI fan.
The system it self passes Prime95 and seems fine, its just the RAID portion. I am not sure if its the drives, the Case's backplane, *Raid card* or Drivers.
I am running out of time on a return so I am probably going to buy a different RAID card (I can't wait on Highpoints RMA proccess).
I am looking at the following as replacements:
areca ARC-1280ML-2G (this has been stated as the most stable with Sata drives but the 7K2000 isn't listed in arecas HCL)
areca ARC-1680IX-24-2G
areca ARC-1880ix-24-4G (This is a bit out of my price range)
I want something that will just work at this point. I will continue scavenging the forums for similar builds with the 7K200 drives but I cannot return those at this point.
Any tips or advice is greatly appreciated.
Thanks!
BIOS/Firmware Version: 1.2.15.23
Driver version: 1.2.25.9 (storport)
WebGUI version: 1.0-080120
Operating System: Windows Server 2003 x64 Standard
Motherboard Model: Asus P8P67 PRO
System BIOS version: 0701
System memory size: 8GB
HDD Model: Hitachi HDS722020ALA330
HDD Firmware version: _______
RAID Configuration: 3 Raid 5s with 8 drives in each
Case: NORCO RPC-4224 4U
Motherboard: ASUS P8P67 PRO
PSU: CORSAIR CMPSU-850TX 850W
Raid: RocketRaid 3560
Other configuration information:
When using the VSS driver (1.2.28.28) Raid drives would disconnect from windows and EVENT was logged:
Source: hptiop Event ID: 9
The device, \Device\Scsi\hptop1, did not respond within the timeout period
This Message will repeat about 2 times and then once or twice later on with 20mins in b/w. I can still login to windows and kind of use it(i can access the C: drive and pull up event manager but anything else will freeze it including trying to shutdown.
When using 1.2.25.9 driver, entire server locks up and events are made every 30Seconds until forced reboot:
source: hptiop Event:129
Reset to device, \Device\RaidPort0, was issued.
With this driver I can do nothing at all on the server, complete freeze the second it happens.
I can still connect to the out-of-band managagemnt webpage for the raid card and no events are logged on it at all. If I submit a rescan on the drives the webpage will then lock up and I can no longer get to it.
I have 2 RocketRaid2320 and they never had issues with these drives.
I have tried switching PCI slots. swaping those 2 drivers, spreading the drives power across the PSUs molex and Sata connectors. Adding the PCI fan.
The system it self passes Prime95 and seems fine, its just the RAID portion. I am not sure if its the drives, the Case's backplane, *Raid card* or Drivers.
I am running out of time on a return so I am probably going to buy a different RAID card (I can't wait on Highpoints RMA proccess).
I am looking at the following as replacements:
areca ARC-1280ML-2G (this has been stated as the most stable with Sata drives but the 7K2000 isn't listed in arecas HCL)
areca ARC-1680IX-24-2G
areca ARC-1880ix-24-4G (This is a bit out of my price range)
I want something that will just work at this point. I will continue scavenging the forums for similar builds with the 7K200 drives but I cannot return those at this point.
Any tips or advice is greatly appreciated.
Thanks!
drescherjm
[H]F Junkie
- Joined
- Nov 19, 2008
- Messages
- 14,933
Usually you can not change brands and keep your arrays.
Usually when your work situation is potentially on the line, one suggestion is to get it resolve fast under a calm environment. Since this is not branded built and looks like you are on your own,
1. Always remember to backup the existing data 1 or 2 copies. I understand you have 24 hard disks so this is not an easy task, figure a way, or build with multiple external enclosures with eSATA connection, don't waste time on USB2.0.
2. If the situation allows and you can personally do this on calm scenario, you can do a stage migration where you have both HighPoint and Areca simultaneously live in the server. Backup portion data from HighPoint to external enclosure/disks with eSATA. Once done, remove that RAID portion from HighPoint, reconnect them to Areca, then build array in Areca. when ready, move data from external enclosure/disks back to Areca Array. Repeat this procedure as many times as possible.
3. To minimize problem for you under potential demanding situation, once you got the card install and operating on server, provide exact model details and other info and post it here. I believe there are many Areca users in the forum, with exact info, they can give good suggestion to make sure everything is taken care before you start the migration. You don't want in the middle of migration you discover issue on the Areca due to non-optimal setup. I do not have Areca card so I cannot comment further.
4. Plan ahead as copying data from 24-disks could take a really long time.
5. As noted, Areca, LSI, Adaptec, HP, IBM and probably Intel all have high-end RAID cards.
1. Always remember to backup the existing data 1 or 2 copies. I understand you have 24 hard disks so this is not an easy task, figure a way, or build with multiple external enclosures with eSATA connection, don't waste time on USB2.0.
2. If the situation allows and you can personally do this on calm scenario, you can do a stage migration where you have both HighPoint and Areca simultaneously live in the server. Backup portion data from HighPoint to external enclosure/disks with eSATA. Once done, remove that RAID portion from HighPoint, reconnect them to Areca, then build array in Areca. when ready, move data from external enclosure/disks back to Areca Array. Repeat this procedure as many times as possible.
3. To minimize problem for you under potential demanding situation, once you got the card install and operating on server, provide exact model details and other info and post it here. I believe there are many Areca users in the forum, with exact info, they can give good suggestion to make sure everything is taken care before you start the migration. You don't want in the middle of migration you discover issue on the Areca due to non-optimal setup. I do not have Areca card so I cannot comment further.
4. Plan ahead as copying data from 24-disks could take a really long time.
5. As noted, Areca, LSI, Adaptec, HP, IBM and probably Intel all have high-end RAID cards.
Last edited:
one year 2009 thread with similar setup but running Fedora
http://hardforum.com/showthread.php?t=1416399
I forgot about Fedora Live. If you have Fedora 14 LiveCD/USB or latest Ubuntu LiveCD (I am not sure about drivers, hope they include HighPoint driver for this setup), you can boot with them.
1. The purpose of that is not to ask you to change OS. It is to verify whether software driver is an issue. If Linux boots OK and can mount the RAID-logical-disks and filesystems, test copy tens of terabytes multiple times between RAID groups so as to exercise all the 24-disks simulataneous forcing it to high workload. This will see whether the underlining hardware is stable or not when all the disks are in motion.
2. If point 1 OK, then it is mostly Windows driver issue. If point 1 is NOT ok, either you have same problem with Linux drivers, or you may need to dig deeper into hardware setup, potentially something more subtle. I remember odditory or someone else last talked about modifying Areca controller timing per getting a response from the drives. I am not sure whether it is relevant or you can indeed do that with HighPoint cards.
3. Finally, to help deduce your power issue, can you disconnect 16 drives and see whether you have the same issue ? how about disconnect 8-drives and re-run the tests? From spec I understand TX850 should be enough. Since you are having problem now, it is best to work through every potential area with open mind. List out every parts installed inside the server including fans.
http://hardforum.com/showthread.php?t=1416399
I forgot about Fedora Live. If you have Fedora 14 LiveCD/USB or latest Ubuntu LiveCD (I am not sure about drivers, hope they include HighPoint driver for this setup), you can boot with them.
1. The purpose of that is not to ask you to change OS. It is to verify whether software driver is an issue. If Linux boots OK and can mount the RAID-logical-disks and filesystems, test copy tens of terabytes multiple times between RAID groups so as to exercise all the 24-disks simulataneous forcing it to high workload. This will see whether the underlining hardware is stable or not when all the disks are in motion.
2. If point 1 OK, then it is mostly Windows driver issue. If point 1 is NOT ok, either you have same problem with Linux drivers, or you may need to dig deeper into hardware setup, potentially something more subtle. I remember odditory or someone else last talked about modifying Areca controller timing per getting a response from the drives. I am not sure whether it is relevant or you can indeed do that with HighPoint cards.
3. Finally, to help deduce your power issue, can you disconnect 16 drives and see whether you have the same issue ? how about disconnect 8-drives and re-run the tests? From spec I understand TX850 should be enough. Since you are having problem now, it is best to work through every potential area with open mind. List out every parts installed inside the server including fans.
Thanks for your replies.
New card showed up: areca ARC-1680IX-24-2G
but it will not read the Hitachi drives. it will not even spin them up. On my norco 4224 i only have green lights for those drives, they never flicker the blue data light. I can put in a Samsung drive and it will stay solid green(power) and blue(data) but it does not appear to spin up either. Neither will be detected by the raid card. I am at a loss of what to do next. I have the areca connected to the norca backplane with the SAS 8087 cables I will try the straight to sata cables as a temp to see what I can get. http://www.newegg.com/Product/Product.aspx?Item=N82E16816133034&cm_re=SAS_8087-_-16-133-034-_-ProductAny thoughts?
New card showed up: areca ARC-1680IX-24-2G
but it will not read the Hitachi drives. it will not even spin them up. On my norco 4224 i only have green lights for those drives, they never flicker the blue data light. I can put in a Samsung drive and it will stay solid green(power) and blue(data) but it does not appear to spin up either. Neither will be detected by the raid card. I am at a loss of what to do next. I have the areca connected to the norca backplane with the SAS 8087 cables I will try the straight to sata cables as a temp to see what I can get. http://www.newegg.com/Product/Product.aspx?Item=N82E16816133034&cm_re=SAS_8087-_-16-133-034-_-ProductAny thoughts?
Last edited:
drescherjm
[H]F Junkie
- Joined
- Nov 19, 2008
- Messages
- 14,933
Staggered spinup?
Posted via Mobile Device
Posted via Mobile Device
something's clearly being done wrong. can you post photos showing what's plugged into what? also post a screenshot of your areca web management GUI including the event log.
Also, while this doesn't likely have anything to do with your current problem, if you build any more servers for a corporate environment I'd strongly suggest sticking with Supermicro motherboards, or at the least a server class chipset with Xeon class CPU(s). MB's with PCIe slots that were optimized to take multiple video cards seem to exhibit more weird glitches when running I/O cards like raid and hba adapters than boards that were intended for the server market. I'd personally never run a storage server with a gaming motherboard, let alone in a business environment. Something to think about.
Also, while this doesn't likely have anything to do with your current problem, if you build any more servers for a corporate environment I'd strongly suggest sticking with Supermicro motherboards, or at the least a server class chipset with Xeon class CPU(s). MB's with PCIe slots that were optimized to take multiple video cards seem to exhibit more weird glitches when running I/O cards like raid and hba adapters than boards that were intended for the server market. I'd personally never run a storage server with a gaming motherboard, let alone in a business environment. Something to think about.
Last edited:
I appear to be having this exact same problem: http://hardforum.com/showthread.php?t=1555189&highlight=rocketraid+drive Its like the rocketraid is a virus. Moved them back and deleted the raid off but they still won't show up. going to hook up again and see what other options there are.
I've got a simmilar problem. Got the same card and enabled staggered spinup on all the drives. Then went to move the drives to another system and the drives won't spinup. According to WD. This will happen on most consumer models and you cannot disable it even though the card says you did. Same thing with several other brands of drives. I can use the drives individually with any Highpoint card but not on motherboard sata ports. I was told by WD that I would have to RMA the drives to use them in any other computer. Might try setting the staggered spinup on the new card and see if it reconizes the drives then. Your RAID problem sounds like a timeout issue but I don't know for sure. If I recall correctly I read something about needing TLER-Enabled drives or something like that for RAID. Could be what's causing the problem.
Last edited:
I put one of these drives from the rocket raid in 3 differrent PCs. None of them will pick it up, Spinrite will find it as a Dos A: undetermined format, but can't do squat with it. Paragons' can't see it at all. I found the 4 extra Hitachis i ordered with the original 24, ( I use them for off site backups) and the areca sees them fine. I should note that the Samsung came from an older 2320 rocket raid too.
Yes, basically those drives are hosed and will only work in Rocketraid cards now.
If you're interested, I may have a solution for you. It may be that the 3xxx card series are just not compatible with your motherboard. I think the 4xxx card series is more compatible. I have a Highpoint 4320 and a Chenbro SAS expander that I've used with those same Hitachi drives without any problem. With this combination you would have 28 ports if you route the external ports on the sas expander back inside the case. If you're interested in trying this combination out, let me know by PM.
If you're interested, I may have a solution for you. It may be that the 3xxx card series are just not compatible with your motherboard. I think the 4xxx card series is more compatible. I have a Highpoint 4320 and a Chenbro SAS expander that I've used with those same Hitachi drives without any problem. With this combination you would have 28 ports if you route the external ports on the sas expander back inside the case. If you're interested in trying this combination out, let me know by PM.
Thank you very much for your rapid responses. Would have been a long night otherwise.
When I disabled the staggered settings I did not have all the drives plugged in, after adding them back to the 3560 did not resend its new disabled settings to them. I had to turn the setting back on, then off again. Hope this helps others out.
With the 3560 - I should note that I disabled NCQ on 16 of the 24 drives and the server went nearly 3 days with out crashing; but I had already ordered the new areca so I am replacing it regardless. Disabling the NCQ could have been the fix, but I had limited time with my return policy to let it go longer than the 3 days.
I will also look into supermicro's motherboards. May be able to take advantage of the SandyBridge chipset recall for that; I don't use the 3.0 Sata but I do use the ESata.
Thank you all so much!
When I disabled the staggered settings I did not have all the drives plugged in, after adding them back to the 3560 did not resend its new disabled settings to them. I had to turn the setting back on, then off again. Hope this helps others out.
With the 3560 - I should note that I disabled NCQ on 16 of the 24 drives and the server went nearly 3 days with out crashing; but I had already ordered the new areca so I am replacing it regardless. Disabling the NCQ could have been the fix, but I had limited time with my return policy to let it go longer than the 3 days.
I will also look into supermicro's motherboards. May be able to take advantage of the SandyBridge chipset recall for that; I don't use the 3.0 Sata but I do use the ESata.
Thank you all so much!
Holy mother of all that is good and gravy! I put the drives back on the RocketRaid and then disabled staggered drive setting. Put them on teh areca and now they show up!!!!!!!!!!! F U RocketRaid F U in the A!
Glad to see you got them all working again. I, on the other hand am still screwed as mine are out of warrenty and cannot RMA them. So out of 24 2TB Drives, 11 are WD and can't be moved. Glad I started going to HITACHI.
Richard
PS: Anyone want to buy some WD 2TB Drives
Just kidding. I'm actually still using them.