Just wondering if anybody has been seeing an abnormally high occurrence of drive failures on PV220s coupled with Perc4/di cards? We purchased six PV220s and put them into production less than 2 months ago, and 4 of them have had drives dropping out and spontaneous array re-builds, on more than one occasion several drives simultaneously dropping out. So far 12 drives or so out of about 100 and counting in 2 months of production. Sometimes its corrupted databases with syslogs indicating hardware failures, but array manager saying all ok. All atmospherics and power is fine. Firmware is the latest all around. Servers are Dell 2850s. Dell support is unable to help.
This is a little off topic, but if you have a need for that much storage space (6 PV220S units), maybe you should consider looking into buying a SAN (Dell|EMC CX500 or so).
We had problems like this and it was traced to a bad back plane on the SCSI controller, is it normally the same drive slots that die? Are you using Seagate drives?
The organisation I work for have a number of 2850's connected to 220's using both the PERC4e/DC (PCIe) and PERC4/DC (PCIX) cards. We have had a number off serious issues with RAID5 + 1 hot spare volumes being corrupted due to failure of HDDs. We primarily use split backplane 220's and utilise 146GB HDDs. We have used both configrations of 15K drive speeds and 10K drive speeds (but not mixed togehter).
On 1 system we have now had more than 4 issues with drive failures causing loss of RAID5 arrays. We have had dell techs on site 3 times trying to fix this issue. The last tech visit, they updated our server to the latest and greatest firmwares and driver combinations. Dell have also replaced the entire 220 with HDDs, backplanes, cables as well as everything in the 2850 apart from the physical chasis, power supplies and memory modules.
So in short, yes, I have noticed a high number of drive failures combined with the 2850's/PERC4 cards.
We have many 2650's with the older PERC3 cards and have not had these issues, although we have experienced the odd drive failure.
Having the same problem with our 2850 & PV220s. It worked fine on the perc4e/dc when the PV220s was in joined bus mode with raid 5 across the whole cabinet with one hotspare, but now It's dropped two different drives in this new mode. The current setup is the perc4e/di is mirrored, the PV220s is in split mode with half of the cabinet on each channel of the perc4e/dc with drive 0 & 15 as hot spares running RAID 50. Drive 5 is the latest one that has failed. I checked in the Card bios and it reports no media errors. The drives are 10k Seagates ST3146807LC REV DS09. I have the latest firmware on the Card, drives & enclosure. I'm also running the latest drivers & latest Dell Openmanage.
Later,
Sean
Message Edited by seanwales on 09-14-2005 03:46 PM
I sent two of my drives to Ultratec Limited in the UK and they said both drives has a SMART trip. Some problems over the buss, however using just the drives they could not pin point why these drives had failed....... It was not the drive but the rest of the hardware that caused the faul. Download Seagate Enterprise and run a full DST test on the failures ;)
"Download Seagate Enterprise and run a full DST test on the failures."
I have had multiple scsi drives fail in arrays, then tested with the Enterprise util... they pass running it multiple times, then fail once replaced into an array. Replaced those particular drives with new retail drives, the arrays are running without problems. Pretty useless test.
As a note, I lost an FSMO raid 5 array by flashing a drive firmware with the util. Works on raid 1 drives though.
"Paul,
Normally a firmware upgrade does not cause any issues. Unfortunately in your case it seems to have caused more problems than it fixed. You can contact our warranty department for information on getting the drives replaced. There phone # is 1-800-468-3472 or you can set up a replacement through our website @ http://support.seagate.com/customer/warranty_validation.jsp
Robert H.
Seagate Technical Support"
"PERC4e/DC seems to be a cause according to Dell.
They suggest that replace PERC4e/DC with PERC4/DC."
Strange, I have multiple client's with this adapter, no problems, and on multiple forums there are no complaint's... with the adapters connected to standard backplanes. Lsilogic must be purposely adding code to the adapter bios to mess with the 220s. they must have some pretty nasty people working for them :smileyhappy:
Message Edited by pcmeiners on 09-29-2005 02:02 PM
Have now had 4 of 6 PowerEdge 2850s with 220s configurations fail and am now up to 8 failures in all on these 4 systems. Most of the configurations were RAID5 but the last one was a RAID10 on 10 disks with 2 hot spares.
Pretty frustrating for me and Dell! :smileymad:
Dell have suggested that I might like to try rebuilding the entire system from scratch. Initialze all disks and rebuild the OS and restore the data. I told them that this would be nice but is not practical. The 6 systems include 111 Disks in total (almost 14.5TBs).
They have also suggested that it may be a problem only with the PERC4e/DCs (i.e. PCIe cards not the PCIX cards). I have not tested this.
Dev Mgr
6 Operator
•
9308 Posts
•
3 Points
622
0
Posted September 8th, 2005 11:00