I've not actually seen this bug with my own eyes on the PanP5. It seems to only hit our older Servals (no longer in production).I've gone through the above linked bug report regarding this issue in preparation for the imminent arrival of my PanP5.
Pyzome, did *you* have it?
Hope you guys don't mind me chiming in... When I got my PanP5 laptop about 2 months ago it did have the dreaded 2.6.28-11 kernel preloaded, but the first updates it ran moved me to -12 and I think I'm at -13 or -14 now (at work so I don't have laptop booted up right now) but I'm on the latest in the Ubuntu repositories.
When I ordered my laptop my fear was that I'd get this bug, but I haven't had any of the corrupt file system problems seen with this bug (knocking on wood) and I've copied MANY large files onto the laptop.
Take care -
Sam
I can't recall offhand if I used ext4 or ext3; I'll have to run fsck from the livecd so I can fix it enough to tell. I *do* know it uses kernel 2.6.31.10, which should be well past the fix.
I had this same problem before with both ext3 and ext4 in jaunty, so I'm not convinced it's that, but I'll take a look in the next couple days.
For what it's worth, I am also seeing the problem with 2.6.31.10 on Ext3. For some reason, it seems to manifest itself primarily with bzip2 when unzipping a large (>100MB) file.
BTW, I am not running System76 hardware.
I haven't noticed an association with large file transfers; I suppose it depends on how you define large. My primary work involves use and conversion (to OO.org) of Microsoft Office documents, often 10-25mb powerpoint files, and pdf files up to 50mb. I transfer many small files frequently and download frequently (the office files mentioned above), but moving clumps of files exceeding 1-2 gb is rare; might have done it a couple times since the install.
I boot and shut down frequently, I *think* I was using the closed Nvidia drivers, but otherwise don't mess with much except the interface layout. I mention the OpenOffice powerpoint->impress because impress files are simply zipped collections; I don't know what OO.org uses to decompress them, but maybe it has to do with your bzip problem, EmmEff
Well, I recently had the opportunity to upgrade my system again, this time laying down an encrypted partition and putting Karmic into it. I found that the corruption bug is still there! I was unable to replicate it reliably, even using the iozone filesystem benchmarking tool, but it would sometimes appear within 10 minutes.
In any event, I found a setting in my BIOS to set the OS compatibility mode from "Other" to "Vista", which shows another field that lets you switch between IDE and AHCI mode. I used this switch and my system has been running fine!
Caveat: I had to scrap my Windows XP partition to do this, and installed Windows 7 in its place. Also be aware that changing this will likely screw up any existing Windows installation of any version.
Bookmarks