Date: Wed, 24 Jun 1998 20:39:12 +0000 (GMT) From: Terry Lambert <tlambert@primenet.com> To: mrcpu@internetcds.com (Jaye Mathisen) Cc: tlambert@primenet.com, hackers@FreeBSD.ORG Subject: Re: NFS send error 55 for fs: Message-ID: <199806242039.NAA13931@usr04.primenet.com> In-Reply-To: <Pine.NEB.3.95.980624030635.513B-100000@schizo.cdsnet.net> from "Jaye Mathisen" at Jun 24, 98 03:08:19 am
next in thread | previous in thread | raw e-mail | index | archive | help
> I notice that it's specifically tied to INN, if I shut off innd, I > don't get the error, regardless of the combo's of diskex, iozone, and > bonnie I run. > > Turn on INN, wait a few minutes, and shazam. I'm betting that you have mmap() enabled in INN. I have been able to demonstrate more serious errors than this using 2.5.x and 2.6.x and the mmap in the Berkeley db code -- specifically, I have been able to get crontab to be corrupted while I was accessing a Berkeley db at the time cron was running. Apparently, the page get reused, but not dissociated from the VM object (this appears the most likely explanation). When the data is written back to the page from user space (in the db code), the page that was being used as backing for the crontab data is corrupted (ie: the page was reused for crontab). This is intermittently repeatable. I don't know, yet, whether it affects 3.x or not. My gut feeling is that it does, despite your short testing. I would *strongly* recommend that you avoid using mmap() at this time, and certainly avoid using it in FreeBSD 2.x. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199806242039.NAA13931>