Page 1 of 2 1 2 LastLast
Results 1 to 20 of 27

Thread: Invalid PEOF result in DFA w/Cheetah RAID

  1. #1
    Join Date
    Oct 2000
    Location
    Colorado Springs, CO
    Posts
    936

    Default

    "Invalid PEOF" followed by some numbers...what's it mean?

    OK. Here's the deal. I get the above error message when running DFA on my internal Cheetah RAID running off of a Miles2 card...but only after a few months of perfect operation. It is always fixed via DiskWarrior or the like, but the simple fact that the error message pops up EVERYTIME without fail...simply irritates me. Here's some vital specs, and yes...I am up to date on all the OS-related goodies (non-beta).

    PowerMac 9600 w/XLR8 300 G3/1MB
    768 MB 60ns interleaved RAM (bought/installed all at once)
    Initio Miles2 card in slot 1
    OrangeMicro FireWire/USB in slot 2
    Voodoo3 200 in slot 4
    ATI NexusGA 16MB in slot 5
    Farallon 10/100 NIC in slot 6

    I have tried basic PCI repopualtion...initialized the RAID (used SoftRAID since day 1)... tons of other goodies. I have been using Macs since 1985 and reapiring them professionally for 3 years. The RAID itself is a pair of ST39102LW's with nothing else on the bus. The factory SCSI bus contains a Plextor Ultra40, a Yamaha 8824 burner and a 16GB Barracuda internally. All termination is A+ guaranteed on both buses, the LVD being handled by Granite cabling/termination. I have initialized the RAID twice in the past year, and follow a regimented monthly barrage of DiskWarrior/SpeedTools/DFA/DiskWarrior with almost no errors until a few months pass. While my system has only required a fresh install once (Other than after initialization) in a year and a half (barring system upgrades), this issue troubles me. Is it an indicator of something larger hidden somewhere?

    And last but not least...yes, I do backup. But I was just looking for some simple reassurance that all is well and DFA has its shortcomings. Thanks for the forum...long-time reader, first-time poster.



    ------------------
    "The first step in attaining knowledge is admitting you know nothing."

  2. #2
    Join Date
    May 2000
    Location
    wherever I hang my hat
    Posts
    3,575

    Default

    I believe that an invalid PEOF refers to a "premature End of File," which means that the Mac is confused about where a file ends. Remember that one thing the Finder does is keep track of where files begin, and where the actual data resides. The Mac will attempt to write files to contiguous blocks. If your drive is fragemented, or if you do not have enough free contiguous blocks, then the Mac will insert a pointer to another block, and resume writing the file there. This results in your files occupying disjointed, non-contiguous blocks on the drive media. If your Mac gets confused about where the files begin and end, you can assume a couple of things:

    1. you need to rebuild your desktop.

    2. you need to run Disk Doctor 5.03 and make sure that you are getting clean passes.

    3. you need to ensure that you are running the latest revision of your disk drivers.

    4. you need to ensure that your firmware on your SCSI host is up to date.

    in this case, I would ensure that all data is backed up, then you need to confirm that you are running firmware 1.06b on the Miles2, and SoftRAID revision 2.2.2. Separately update drivers on each drive in the array.

    once this is done, I would use FileBuddy 5.3.8 to rebuild the desktops on all volumes. Finally, ensure that you are using Disk First Aid 8.5.5 or later.



    [This message has been edited by magician (edited 30 October 2000).]

  3. #3
    Join Date
    Oct 2000
    Location
    Colorado Springs, CO
    Posts
    936

    Default

    Magician,

    Thanks for the reply. Rest assured I am indeed using DFA 8.5.5 and MacOS 9.0.4 and have Initio's 1.06b on the Miles2. I also am running SR 2.2.2 and usually cannot remedy this message with any kind of DT rebuild. This time it is esp distressing as I recently had to reformat after a serious loss of integrity with the array. It was built by me, and is my 3rd array under my belt. I have built 2 since then, and this is the only time I have seen this. This time, it has only been less than 3 weeks since a clean reformat/driver install. I have very few apps on the RAID as of yet, as I was testing to see if this would happen. Basically, I only installed my internet apps and updated system components. Most of my other apps I have been running from the Barracuda. DD 5.0.3 reports clean passes...and all other tests I throw at my baby result in flawless peformance. Until this error crops up, and if I leave it go (like I did last time)it ends up taking down my system. I had to try everything. If I repair it...it comes back. if I don't, my system begins random crashing. I have even tried backing down to stock G3 card speed (I am not on the bleeding edge, anyways).

    Keep those suggestions coming. Since this isn't a crucial issue that kills things immediately (and I can fix it when it happens), I am merely looking for an explanation. I hate to think that there is some problem with year-old NOS Cheetahs and a brand new Miles in an otherwise rock-solid lady of a machine.

    Oh, and I found the same definition for the "PEOF" reference. Thanks again!!!

    ------------------
    "The first step in attaining knowledge is admitting you know nothing."

  4. #4
    Join Date
    May 2000
    Location
    wherever I hang my hat
    Posts
    3,575

    Default

    I would definitely rule out an extension conflict of some kind....and check for virii. Use Norton AntiVirus 6.02 with up-to-date definitions.

    something is causing this. It is not a trivial thing that I would ignore. My suspicion is that it is nothing less than corrupted data, and will over time result in more corrupted data.

  5. #5
    Join Date
    May 2000
    Location
    Cordova, TN
    Posts
    2,428

    Default

    Sometimes DD will allow you do make an alias to a corrupted file. It's really cool when it's an invisible file. I suppose you are not that lucky, since this one is not really corrupted, but part of it is lost.

    I've seen Stuffit make a call like that when it tries to expand and/or decode an incomplete download or some FTP site's screw up. I guess, though, if you stored one like that, still compressed, the system wouldn't mind. You might try Fast Find and look for compressed/encoded files.

    Just for kicks, use TechTool 1.2.1 to analyze your system. It can also do a much better job of rebuilding Desk Top DF and Desk Top DB. than the keyboard.

    As you can tell, I'm groping here. These are all WAG's

    [This message has been edited by Louie (edited 30 October 2000).]

  6. #6
    Join Date
    Oct 2000
    Location
    Colorado Springs, CO
    Posts
    936

    Default

    Thanks for the replies, folks....but a couple of thoughts back at ya:

    I have switched ISP's since this has happened, and gone from modem to DSL with a HW Firewall, and it has still happened. The last time I initialized everything and left just the array in (No other HD's, etc.) was this past time, and it has still happened. I have been fine as far as all virus checks with (gasp) Virex and NAV 6. I am truly baffled, now. I just bought NAV 6 and got all up-to-date, and also ran TT on this machine (from a friend...for giggles). Everything turns up A-OK as far as their own checks this time. I have backup of my critical data, but I am still at a loss. Thanks again.

  7. #7
    Join Date
    Jun 2000
    Posts
    769

    Default

    Hey D,

    Two possibilities come to mind, just shots in the dark really. One, possible marginal block on your drive in one of the directory trees (the same locations get reserved for these areas with new formats). Two, possible bad or marginal RAM stick. The diagnostic process would be similar for either possibility, i.e. run stringent tests for one or more days continuously. FWB Hard Disk Toolkit has a disk test function that is nice with selectable levels of stringency. TechTool Pro has a nice cluster of RAM tests.

  8. #8
    Join Date
    Oct 2000
    Location
    Colorado Springs, CO
    Posts
    936

    Default

    MM,

    Will do on both accounts. I will leave one test cranking all night tonight and one all night tomorrow (unless something crops up). I had tested the RAM for 48 hours straight (anal? yes.) when I got it and then again about 3 months later. That was about 6 months ago. Thanks for the shots in the dark that may just help shed a little light.

  9. #9
    Join Date
    Oct 2000
    Location
    Colorado Springs, CO
    Posts
    936

    Default

    OK. Here's the new deal. All tests passed with flying colors. I ran the HDT test on my HD's for 51 hours straight with A-OK test results. RAM test initially froze after about an hour and 50 minutes. But upon restart, I successfull ran the test for two separate 1-day tests. So now I am where I was to begin with. Umm...OK. That's all for now. I will post if there is any new info, or maybe the 45-day radar-scope for posts will let this one go unanswered and fall between the cracks. Thanks to all for their suggestions and efforts thus far. I am wiser and better suited for any future problems, I just wish my current one had an answer. Maybe it's a sign to ditch the setup and finally get that G4 with a PCI chassis outboard...hmm....

    ------------------
    "The first step in attaining knowledge is admitting you know nothing."

  10. #10
    Join Date
    Jun 2000
    Posts
    769

    Default

    Hey D,

    You mentioned a "serious loss of integrity" with your RAID a while back--what happened back then?
    Also, what optional settings are you using for the SoftRaid driver?

    I have never seen a troubleshooting thread on this site drop off the radar scope unless the original poster faded away

  11. #11
    Join Date
    Jun 2000
    Posts
    769

    Default

    Hey D,

    Just thought of something else that may give a clue. Try running MacsBug and if you do not drop into it with a bomb or freeze, manually drop into it when you get the error message.
    Do a "stdlog" to see what's going on. I am NOT a programmer and don't have that much to offer here but sometimes the results are clear enough to understand, even for me. Maybe someone else with a deeper understanding can suggest a better way to use MacBug.

  12. #12
    Join Date
    Oct 2000
    Location
    Colorado Springs, CO
    Posts
    936

    Default

    And the crowd goes wild!!! So I had totally forgotten in all the mayhem about seeing what MacsBug had to say about all this. Thanks for the reminder.

    As far as my loss of integrity, well...I crashed in the middle of things (typing into PageMaker 6.5) and then whammo!! I tried a restart to a flashing "?"..I booted from OS9's CD and the drives were recognized, but there was no volume. I booted from just about every bootable Mac CD I think exists and tried most things short of what Darin went through with data recovery specialists. Turns out I had to initialize/reformat/reinstall drivers (not in that order) for the array. In the meantime, I could boot perfectly off of my CD drive and the other internal drive just fine. I received an "Invalid sibling extent" message (or something to that effect) through DFA. I lost over 200 pages of a manuscript, of which most was backed up. But the point is, it all started with that same error. Invalid PEOF. But that time was the one time I let it go to see just how far it would go. I usually remedied the problem through a utility other than DFA, but I had to see how bad things could get. Mind you, the system itself showed no unexpected crashes or slowness between the initial instance of the error report and the bad crash. It took less than 2 weeks for it to destroy the stripe. Or maybe it was totally unrelated, but I doubt it. Thanks for the help, and when I get something interesting in MacsBug, I will post...or when I get something intersting period.

    Seriously, thanks for all the help. I am so glad to see this kinda dedication and interest coming from volunteers/addicts like yourselves.

    ------------------
    "The first step in attaining knowledge is admitting you know nothing."

  13. #13
    Join Date
    Oct 2000
    Location
    Colorado Springs, CO
    Posts
    936

    Default

    One last thing, MM. As far as optional settings in SoftRAID? Well, I was schooled in the set it & forget it scholl of SoftRAID. So I left things as is. The buffer settings are both at "0", the write cache is enables, and there are 3 cache segments. If these are not the settings you were referring to, please point me in the right direction. That is all for now. Live long and prosper.

  14. #14
    Join Date
    May 2000
    Location
    wherever I hang my hat
    Posts
    3,575

    Default

    I still think it is a driver problem....it almost has to be an extension which is doing this. I would simplify dramatically, and prune extensions without mercy. If it is not mainstream, nuke it. Hell, if it isn't installed by MacOS, nuke it.

  15. #15
    Join Date
    Oct 2000
    Location
    Colorado Springs, CO
    Posts
    936

    Default

    Mag,

    While I think it is most likely what you are suspecting....I have a ton of 3rd party extensions that (professionally) I cannot Nuke. I have only what I need on this system right now. Mac OS9 and all its pertinent updates...Photoshop, Pagemaker, Premiere, HP printer, Wacom tablet, ATI and 3dfx drivers, CD/DVD SpeedTools and Explorer. Seriously..that is it. No nothing else. If those add-ins are causing this...then it is not a SW problem, cuz I have the EXACT setup on a VERY similar 9600 at my office that has never had these problems. No, I have not tried swapping drives. But I have tried everything other than that. Thanks for the conitued support. If you need a system profile or anything, lemme know and I will email you one.

    Thanks again and again.

  16. #16
    Join Date
    Jun 2000
    Posts
    769

    Default

    Hey D,

    Yeah, those were the settings I was asking about. Also, since you use defaults, I assume you are using the default stripe allocation block size, no photoshop acceleration and the standard memory pool setting.

    Just to complete the survey of likely hardware issues, I read that your cabling and termination are verified OK. As Mag says---on to software/driver issues. For some reason that is completely unclear to me, I would suspect the Farallon card/driver stuff as a gut-level reaction. Can you try running with the software disabled and/or the card pulled to see if the error occurs? Are you using the latest pertinent ATI extensions?

    'later

  17. #17
    Join Date
    Oct 2000
    Location
    Colorado Springs, CO
    Posts
    936

    Default

    MM,

    I left the block size as is and PS accel off.

    I had my reservations about the Farallon card, too. So I have ordered 2 other-brand 10/100 cards to try things out. They are cheap enough. I can always "eBay" them. I can't let go of my 100base-Tx just yet, as I have a network-dependent machine. I have a couple other Macs running on the network, and I would absolutely hate to step down to 10base-T. I know, I know. If this is the root of the problem, wouldn't it be a worthy sacrifice? Yes. Maybe I will do it regardless. I just hate to use the stock port for my combo DSL/network bandwidth. We will see. I will gladly keep you all posted. And yes, I am up to date on the ATI situation...and just got the 1.1 Voodoo drivers, too. I follow directions exceedingly meticulously, so I am sure I have cicumvented any weird OpenGL/ATI issues, as well. Thanks a thousand times. I love gurus.

    ------------------
    "The first step in attaining knowledge is admitting you know nothing."

  18. #18
    Join Date
    May 2000
    Location
    wherever I hang my hat
    Posts
    3,575

    Default

    it just occurred to me to ask.....did you ever run Disk Doctor or TechTool Pro or anything else and do a media check? It takes time....but since a PEOF is involved, that is one thing that could also cause this. You could have bad blocks on the drive.

  19. #19
    Join Date
    Oct 2000
    Location
    Colorado Springs, CO
    Posts
    936

    Default

    Magi,

    To tell ya the truth? Yes I did. Twice on the array. Once when I first got the drives...and again after the problem began to crop up (a few months ago). Since then? No. Saw no need as I figured that if I passed once and problems still happened, it was most likely not the cause.

  20. #20
    Join Date
    May 2000
    Location
    wherever I hang my hat
    Posts
    3,575

    Default

    well....heat damage can cause the media to buckle...which could have damaged media...though you would also think heads would be damaged, too.

    might be something to try. Something is certainly amiss.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •