Announcement

Collapse
No announcement yet.

File coruption on Raid array with IBM GXP75's!

Collapse
X
Collapse
 
  • Page of 3
  • Filter
  • Time
  • Show
Clear All
new posts
Previous 1 2 3 template Next
  • #1

    File coruption on Raid array with IBM GXP75's!

    As some of you might remember Two of my three IBM GXP75 30GB drives got bad sectors and I RMA'd them.

    What I got in return was two "Repaired" drives....

    I use them in the Raid.

    And emidiatly I got the Usual paranoid feeling that something was not right.

    And Now I have File coruption errors....

    To date its 6 filed that I have detected to be bad (after checking 400 by hand, URGH...)

    Three i could recover from yesterdays backup (yes I have started to do nightly backups to my NLE machines HDD) the rest I have to redownload.

    And tomorow I will call IBM suport and be very very nasty!

    I vill demand new ones and quick!

    It's iritationg to be hanging with Samsungs for a month again.
    The two ass slow Samsungs I had never did this kind of things and they I had for a month.

    This is on a promise onboard RAID!

    And the samsungs never did this!

    and the funny thing is that it's not big files either.

    Could it be the shutdown bug? (ie coruppted files because of cahces that has'nt been written to HDD?)
    If there's artificial intelligence, there's bound to be some artificial stupidity.

    Jeremy Clarkson "806 brake horsepower..and that on that limp wrist faerie liquid the Americans call petrol, if you run it on the more explosive jungle juice we have in Europe you'd be getting 850 brake horsepower..."
    Tags: None
  • #2
    Have you considered using another controller (RAID or not) to see if that's the problem? Have you tried disabling fast shutdown?
    <TABLE BGCOLOR=Red><TR><TD><Font-weight="+1"><font COLOR=Black>The world just changed, Sep. 11, 2001</font></Font-weight></TR></TD></TABLE>

    Comment

    • #3
      It would make sense if it did happen after I had visited Win98SE...
      BTW "Fast Shutdown" is Disabled by default in SE.

      I dont have another RAID controller (Its to damn expensive to run out and by one to test.)
      And I put them on the onboard via686B I am destined to get problems according to a lot of persons on the fori... (Just joking )

      The old IBM's dident behave like this, the samsungs dident, the new one does!

      Im downloading IBM's latest Drive fitnes test now.

      Im going to get all the specifik updates to W98se and use the "Sync" utility to flush the caches and see what will happen.

      Worst thing is:

      Last time I removed the array and booted to w2k by mistake, w2k Puked and hosed itself!
      If there's artificial intelligence, there's bound to be some artificial stupidity.

      Jeremy Clarkson "806 brake horsepower..and that on that limp wrist faerie liquid the Americans call petrol, if you run it on the more explosive jungle juice we have in Europe you'd be getting 850 brake horsepower..."

      Comment

      • #4
        Yes, the 686B can cause data corruption. A proven fact.

        IBM had a bad run of 60GXP and 75GXP drives starting last fall so you may have got some of those. StorageReview.com's forum is just loaded with posts about them.

        Fortunately mine are older than that and haven't exhibited any problems (knock on wood).

        Dr. Mordrid
        Dr. Mordrid
        ----------------------------
        An elephant is a mouse built to government specifications.

        I carry a gun because I can't throw a rock 1,250 fps

        Comment

        • #5
          Now that everyone likes to the blame the soundcard you don't happen to have a sblive by any chance in there?
          Chief Lemon Buyer no more Linux sucks but not as much
          Weather nut and sad git.

          My Weather Page

          Comment

          • #6
            Fast shutdown is <B>not</B> the same thing as windows shutting down too fast. Go to Windows Update. I have SE and they still have a fast drive patch.
            Gigabyte P35-DS3L with a Q6600, 2GB Kingston HyperX (after *3* bad pairs of Crucial Ballistix 1066), Galaxy 8800GT 512MB, SB X-Fi, some drives, and a Dell 2005fpw. Running WinXP.

            Comment

            • #7
              Yes, I have a SBLIVE.
              No, the drives are on a Promise onboard Raid controller.
              Even those ViaHardware or what the site is called (there is a bunch with variations on the same combination of words or similar sounding ) actualy claimes that the corruption was only when the harddrives was conected to the second IDE controller on the onboard 686B northbridge!

              And by the fact that I found a couple of "new" corupted files everythime i checked I would have noticed it erlieer if it not was my "new" repaired august 2001 gxp75!!!

              Now I have moved the files back to my old (and never left my case) IBM drive.

              I have also installed the fast drive patch for w9x!
              If there's artificial intelligence, there's bound to be some artificial stupidity.

              Jeremy Clarkson "806 brake horsepower..and that on that limp wrist faerie liquid the Americans call petrol, if you run it on the more explosive jungle juice we have in Europe you'd be getting 850 brake horsepower..."

              Comment

              • #8
                To be honest data corruption can also occurs when hard drives are connected to the same channel. I've seen this happen plenty of times. Also data corruption occurs with other none creative cards. Even better data corruption occurs when there are no soundcards present.
                It these factors which make me believe the problem is more with the 686b than it is with the sblive although some factors of the live make the problem worse.
                Viahardware recently did a test over several motherboards and only managed to get the sblive to corrupt. This however doesn't follow what poeple have reported in the forums who have reported corruption occuring with many different soundcards.
                Abit and Creative are one of the most popular brands and maybe one of the reasons why they get a lot of write ups on the viahardware forum.
                Chief Lemon Buyer no more Linux sucks but not as much
                Weather nut and sad git.

                My Weather Page

                Comment

                • #9
                  Windows shutting down too fast has nothing to do with whether your HD is connected to a Via chip, intel chip, Aladdin chip...etc
                  Gigabyte P35-DS3L with a Q6600, 2GB Kingston HyperX (after *3* bad pairs of Crucial Ballistix 1066), Galaxy 8800GT 512MB, SB X-Fi, some drives, and a Dell 2005fpw. Running WinXP.

                  Comment

                  • #10
                    I still reckon it's going to be the 686b bug having a little laugth here.
                    Chief Lemon Buyer no more Linux sucks but not as much
                    Weather nut and sad git.

                    My Weather Page

                    Comment

                    • #11
                      Sigh...

                      Lets get some facts straight!

                      I did had IBMs on the Promise cobtroller erlier and they did not have any coruption problems!! (despite the sector errors that hosed my array they still barfed up my data unharmed!!)

                      For a Month I had two Samsungs on the same Promise controller without any coruption!

                      I get two "repaired" Ibm drives and I discover coruption wthin a week!

                      When residing on the boot drive conected to the aledged evil 686B no coruption occures!

                      I made a non compressed zip file and transfered that betven the raid array and the boot drive while using winamp!
                      That did not corupt the file!

                      It seems that only files that are rsiding in the array gets corupted!

                      It don't happen when copying or moving from or to the array!

                      The only other things that changed through that time was that i replased the overklocked duron 700 wit a non overklocked TB1330 and got two 256PC133 CL2-2-2!
                      If there's artificial intelligence, there's bound to be some artificial stupidity.

                      Jeremy Clarkson "806 brake horsepower..and that on that limp wrist faerie liquid the Americans call petrol, if you run it on the more explosive jungle juice we have in Europe you'd be getting 850 brake horsepower..."

                      Comment

                      • #12
                        Hmmm.....

                        I've have 4 60g 75GXP's purchased when they first came out. I'm still using the same 240g array as day one. In fact this array was migrated intact from a Fasttrak100 to the TX4 without any changes at all.

                        I've never had any data corruption of any kind. Same goes for the other four Fasttrak100 arrays in my lab. They are all using MAXTOR drives though, which have been virtually bullet proof here.

                        It could be that you're running into those bad drives I mentioned earlier. Striping them may just bring out a latent issue with the drives themselves.

                        Dr. Mordrid
                        Last edited by Dr Mordrid; 5 September 2001, 10:43.
                        Dr. Mordrid
                        ----------------------------
                        An elephant is a mouse built to government specifications.

                        I carry a gun because I can't throw a rock 1,250 fps

                        Comment

                        • #13
                          The IBM suport mentioned vibrations...
                          Could vibrations cause datacoruption without sector errors?

                          He also said that I could test to lower the UDMA settings from 5 to 4 (100-->66).
                          should I test that? (will do it anyway, but advice is always nice)

                          BTW:

                          He tried to make the impression that it was all the Promise's fault...

                          I don't belive him tough...
                          If there's artificial intelligence, there's bound to be some artificial stupidity.

                          Jeremy Clarkson "806 brake horsepower..and that on that limp wrist faerie liquid the Americans call petrol, if you run it on the more explosive jungle juice we have in Europe you'd be getting 850 brake horsepower..."

                          Comment

                          • #14
                            That's called the "hot potato effect". Suppose you have two devices from hypothetical manufacturers A and B that don't quite work together. When you call A and complain, they will insist it must B's fault and vice versa.

                            *shiver*
                            Someday, we'll look back on this, laugh nervously and change the subject.

                            Comment

                            • #15
                              My only comment on the data corruption is that I went a month without any that suddenly whammo. There was no bios upgrades no new hardware no software patches. True it maybe the new (old) drives but again it could also be the 686b. You could also try reducing you memory timings that what finally solved my data corruption. Hold a minute reading the last bit of your post again "new memory". Also if you read around the data corruption seems to get worse with faster cpu's.
                              Chief Lemon Buyer no more Linux sucks but not as much
                              Weather nut and sad git.

                              My Weather Page

                              Comment

                              Previous 1 2 3 template Next
                              Working...
                              X