RR2322 RAID5 will not init with 8x drives

Sate

Weaksauce
Joined
Jul 8, 2004
Messages
106
A few hours into initializing, the raid controller will report a drive has failed. This only happens when I try to set up a 7 or 8x array. 6x array initializes and has worked fine for a week. Now a number of potential problems I have already investigated as follows:

Bad drive(s)?
After repeatedly trying to rebuild/init the array the "failed" drive rotates at random. The most recently "failed" drive initialized fine then worked perfectly well as part of a 6x array for days.

PSU problem?
The drives are housed in an external array. I brought in a second PSU and split the drives among them in terms of power. This had no effect.. sometimes the "failed" drive was powered by the second 400watt PSU, sometimes the other.

Heat issues?
I reduced the environmental temperature and brought in an extra 200mm fan just to rule this one out. I noted a failure while no drive was more even warm to the touch.

I have no idea what else could the problem.. the controller itself?
All hardware and software is new and otherwise performs flawlessly. There is no overclocking. I put in a support request email to Highpoint but they are of course closed for the holidays, through Jan 5.

System specs
Was WinXP MCE but now Vista-64 Ultimate SP1
AMD Phenom 8450 tri-core 2.1ghz
ASUS MA378-T bios: 0403
2GB Crucial ballistix ram @ 8500
640GB WD HD
ATI Radeon HD 3600 video
RocketRaid 2322, latest bios ver 1.7, Vista driver ver 1.9
RAID hard drives, 8X Samsung 1TB HD103UI drives
SC-ASAT84XB external SATA HD enclosure w/Two Infiniband Multilane 4X 4-channel SFF-8470 external female connectors; has 300watt PSU from factory
 
bad port? is it always the same port that fails? I had 8 drive raid 5 on rr2322 without any issues
 
You've done a good investigation job so far! Like axan said, same cable or port? I've got 16 drives spanned across two RR2220 without issue.
 
bad port? is it always the same port that fails? I had 8 drive raid 5 on rr2322 without any issues

Does not seem to be. The channel listed by the error message changes. Plus early on in trouble shooting I switched the SATA cable from the first "fail" HD with the adjacent disk. On the next run it failed to initialize but cited a third disk.
The error logs in front of me show disk A13957 Controller1/channel4 failed on 12/28
but disk A14022 Controller1/channel8 failed back on 12/24.

The maddening question is, why 6x but not 8x? As I noted before, when I tried to migrate the existing 6x array to 8x.. one of the drives in the old array failed not one of the two ports/disks I was adding to it. If the port is failing, its failing under condition of "more than 6 disks present". :confused:
 
You've done a good investigation job so far! Like axan said, same cable or port? I've got 16 drives spanned across two RR2220 without issue.

Thanks, wish it was getting me somewhere!
Wow 16.. just out of curiosity how do you backup that much data? or is one array the BU for the other?
 
1/2 of the array is filled with 250s, I don't really even use them anymore, need to just pull them out! The important stuff is all backup to separate drives and stored in my fire resistant safe. The REALLY important stuff is again backed up to my detached garage system! Nothing truly "off site" but its as far as I'm going!
 
Update-
I contacted PC Pitstop support and they offered a possible diagnosis that is at least plausible. Some raid controllers do not deal well with 3gb/s mode drive arrays. They suggested using a Samsung disk utility to change the disks to 1.5gbs mode. I tried to do this and ran the utility as prescribed but it did not change the disk's operational mode. Troubleshooting is ongoing..

btw the techs at PC Pitstop are top notch and helpful.
 
you might need the samsung drives attached to a standard SATA controller to change their mode.

does samsung not use a jumper for this as WD and Seagate do?
 
you might need the samsung drives attached to a standard SATA controller to change their mode.

does samsung not use a jumper for this as WD and Seagate do?

That is my working theory. Most Samsung drives apparently do not have a jumper for this. There is a software utility to do so and I have it but I can not use it. This is because you must connect the drives directly to a SATA port on the mobo (not via the SAS/raid) but my ASUS mobo will NOT recognize the drives, period unless I set the onboard SATA to "raid" mode, at which point I still can not access the drives directly.

So, the next move is to get take a drive to an offsite PC that does not have my boards RAID hangup in order to apply the samsung utility.
 
I was able to set the Samsung drives to SATA150 mode using the ESTOOLS utility and an offsite Dell PC. I put the disks back into the SAS and created an array but it failed about 2 hours into initializing again. =/ I have no ideas left.. going to RMA the controller.
 
Update- I replaced the RAID controller with a new one of the same make and model. The result was the same errors. Just for testing I tried creating two RAID5 arrays, one on each bridge. 4x and 4x. This also failed to init.
 
Back
Top