Thanks Simon, but we're pretty sure it's not specifically hardware. It may be HyperV related in one case, but the other box is physical. What we have now seen on both boxes is that the corrupted files are getting what looks like output from a background process which ought to have gone to a COMO or been discarded just written over the top of them.
John Fenlon down here is being very helpful trying to track this down with us. Cheers, Ken On Wednesday, 3 June 2015 18:21:28 UTC+10, Simon wrote: > We have Server 2012 in production but on QM not jBASE. > > This should more likely to be a hardware issue than an OS related > problem. Key area may be RAID disk systems, particularly firmware related > and I've seen this often with the use of write-through caching. I had > problems with DELL PERC-3 systems with iffy firmware where "chatty" files > would corrupt. It appeared that if a simultaneous write and read occured > on the file then they could get out of sync and you would read the data > before the write had taken place! Hashed files such as that used by most > MV systems would get incorrect links to overflow and all sorts.... > > It may not be this at all. > > Having said that, my experience with using jBASE for a no of years (v3 > though) was that I would get "random" file corruptions from time to time > anyways. I always suspected memory management as an underlying cause, but > never pinpointed whether this was jBASE or Windows..... > > Not sure I've helped much! > > Regards > Simon > > On Tuesday, June 2, 2015 at 9:46:17 AM UTC+1, Ken Wallis wrote: >> >> Hi, I'm just wondering how many people actually have production sites >> running jBASE on Windows Server 2012? >> >> We (Paperless Warehousing) have 2 production sites with Server 2012 and >> apart from the horrible mess Microsoft have made of the interface to this >> server OS, everything seems fine except these sites experience at least 10 >> times the frequency of data corruption that we get at any other Windows >> site and that's without them doing random reboots or anything like that. >> >> We have had completely static and never updated files like jBASE's >> internal jbcmessages get corrupted, but most frequently it is an index on >> one of our busiest files that goes south - often somewhere in the early >> hours of the morning. >> >> Of these 2 sites, one is running 5.2.26 and the other 4.1.6.16. The 5.2 >> server seems to have more trouble than the 4.1 site does, but both >> experience far more issues than our Server 2003 or Server 2008 sites do. >> We have been down the path of making sure the boxes aren't being rebooted, >> verifying that the AV tools on the boxes have exceptions for the database >> disks and checking that they aren't using the new ReFS filesystem, but >> still we keep getting files that look like they've had random blocks >> stomped onto them. For example we've had distributed file stubs which >> should be less than 1K suddenly become much larger and stop working - the >> first few bytes seem right, but the part file information is all wrong and >> the file is a few MB. >> >> I'm wondering if anyone has spent enough time with this OS to identify a >> service that Microsoft might have added that thinks it is helping us but is >> actually stuffing us up. >> >> Short of that, I'd just be interested to hear if there are hundreds of >> sites out there running on Server 2012 without any issues - at least I'd >> know it was only us going mad! >> >> Thanks, >> >> Ken Wallis, Paperless Warehousing >> Sydney, Australia >> > -- -- IMPORTANT: T24/Globus posts are no longer accepted on this forum. To post, send email to [email protected] To unsubscribe, send email to [email protected] For more options, visit this group at http://groups.google.com/group/jBASE?hl=en --- You received this message because you are subscribed to the Google Groups "jBASE" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
