ProLiant Servers (ML,DL,SL)
cancel
Showing results for 
Search instead for 
Did you mean: 

DL 380 G3/G4 drive failures

Andy_180
Trusted Contributor

DL 380 G3/G4 drive failures

Has anyone had trouble with multiple drive failures within a few hours in a 380 G4 or G3?
all PSPs are at a respectable level. FW is up to date via FW 7.40 CD. it doesn't seem to matter what RAID level, drive size/speed or stripe size, duplex or simplex.

the drive(s) will be marked as bad until the backplane is replaced. after backplane gets swapped drives are marked as now good but it is too late for the data. Doesn't matter if simplexed or duplexed so that rules out bad terminators.

HP swears they don't have anything on paper about it but i have noticed that the replacement backplanes they send are of a different rev.

we have about 40 DL 380 G3s and G4s and have replaced 5 backplanes in the last 90 days. all servers that dump the backplane also dump the data.

has anyone had the same experience? it seems that of the backplanes that fail, one drive had never been lit. only the activity light would be on. or when it would fail it would turn red. the green cylinder would never light up. never seemed to affect performance or anything.

Thanks in advance!
6 REPLIES
Oleg Koroz
Honored Contributor

Re: DL 380 G3/G4 drive failures

Hi andy
Had you look from prospect of HDD problems?
Is drives identical that you use in G3 or G4?
Had ever-happened same server reported twice failed HDD?
Just for curiosity can you upload ADU report from one of the Server where you experienced Problem?
Andy_180
Trusted Contributor

Re: DL 380 G3/G4 drive failures

Thanks for reply!

I dont think its a drive problem, it happens with 15k, 10k, 146 & 72GB drives.

never happen twice in same server after backplane replaced (yet) (knock on wood). after backplane replaced, "Failed drives marked as replaced and now good. All data lost, please restore data from most recent backup. Hit F1 to continue". Then I cry. then my boss cry.

Here is an ADU from one that went back in DEC. "JAXFPS1_ADU_REPORT". just a fyi... our naming convention is File and Print sharing server in our Jacksonville office. hence JAXFPS1.
Oleg Koroz
Honored Contributor

Re: DL 380 G3/G4 drive failures

Definitely backplane Problem â I hope you get that report before swapping it

BTW firmware for HDD Port 2 SCSI ID 4 Slot 1 Smart Array 642 - Critical

http://h18007.www1.hp.com/support/files/storage/us/download/23632.html


http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?lang=en&cc=us&objectID=c00559846

Backplane part of SCSI bus and sensitive to power fluctuation: which might be cause by one of the drive, or actual backplane problem
Lance Spencer_1
Occasional Visitor

Re: DL 380 G3/G4 drive failures

I support numerous networks that use the DL380 G2/G3 servers and recently have noticed many networks are having SCSI backplane problems. When we go to the remote sight, the fastest way we've used to determine if it's a backplane problem is swapping out the backplane with a known good backplane & see if everything starts working normally. Lately, we've seen a rash of AIT100 tape drives stop working properly & replacing the backplane fixes that problem.

I've been told other technicians have run across this too.

I tried using the utilities on the SMART disk CD (not the most indepth level) and the tests don't come across anything wrong with the hardware. However, changing the backplane has fixed all our problems so far.

Is there a better way of testing only the SCSI backplane to see if it is truely bad without having to swap out backplanes?

I'm supposed to go work on a remote network next week and after talking to the local techs, it seems like a backplane problem again. It'd be nice if I could run a fairly fast utility to identify if these backplanes are bad. (without replacing with known good backplanes)

Lance

Thanks,

Lance
Andy_180
Trusted Contributor

Re: DL 380 G3/G4 drive failures

Thanks Lance. no known utility to diagnose a bad backplane other than a swap. it seems to me one characteristic of a backplane that may fail in the future is that one of the drives (usually scsi id 5, top left)wont have the green cylinder lit. until the backplane fails a drive then the backplane gets replaced. then the green cylinder is there. but that is only a working theory so far. if you look on the original factory backplanes they will be rev. e. if you call it in the one they send you will be a rev. f. the conspiracy theorist within says they know something is up on their back-line support. But of course HP support is playing blind, deaf and dumb yet again on another serious issue.

it scares the doggie doo doo out of me when HPSIM reports a "logical drive status change" in one of our G3/G4 servers anymore.

we have swapped the failed drives with known good drives and they will fail a few hours later. if we are lucky, an affected (or infected) server will only fail one slot. its when all of a suddden 2 or 3 drives fail at once is when we are in trouble.

thanks.
--Andy





Lance Spencer_1
Occasional Visitor

Re: DL 380 G3/G4 drive failures

When I worked on one network, we had to rebuild domain controller 1, two times before we finally replaced the backplane. The indications on this one machine was SCSI drive 1 & 3 would go amber & then server drops offline. We rebuilt the RAID and all software. It stayed online for two days and got the same indication. Reloaded again & it dropped offline (same indication) and then we decided to replace the backplane.

We're coming across tape drives not working now so, we're getting some other indications of bad backplanes. I just don't understand why all of the sudden these things are failing more often.

Lance