Date: Wed, 25 Mar 2009 17:45:38 +0100 From: cpghost <cpghost@cordula.ws> To: "Mikhail T." <mi+thun@aldan.algebra.com> Cc: freebsd-stable@freebsd.org Subject: Re: dump | restore fails: unknown tape header type 1853384566 Message-ID: <20090325164538.GA6843@phenom.cordula.ws> In-Reply-To: <49CA5602.9050001@aldan.algebra.com> References: <49C83673.3000604@aldan.algebra.com> <200903251232.11418.doconnor@gsoft.com.au> <49C99204.2050601@aldan.algebra.com> <200903251334.38350.doconnor@gsoft.com.au> <49C99FD2.50609@aldan.algebra.com> <nospam-1237956961.78753@joker.yaxom.com> <49CA5602.9050001@aldan.algebra.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, Mar 25, 2009 at 12:04:18PM -0400, Mikhail T. wrote: > Generally, when dumping read-write mounted filesystems, one is > supposed to use snapshots, but that is very buggy on its own > (leading to kernel crashes), so it is safer to rely on the FS being > "idle", if it can not be forced into read-only for some other > reason. Perhaps your kernel/world is not recent enough? There used to be problems with snapshots, including hangs or crashes, but IIRC, they disappeared completly a couple of months ago, at least for me. (no, sorry, I can't pinpoint the exact date, and much less the exact revision nr). Snapshots alone, and dump -L + restore (on RELENG_7/amd64) working flawlessly here on 20+ rack-mounted servers; some to tape, some to SAN, but I'm probably just lucky. BTW, that's what makes this problem so intractable: those who could debug it, can't reproduce it on their machines. It's scary to know there's a bug lurking in there. > -mi -cpghost. -- Cordula's Web. http://www.cordula.ws/
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20090325164538.GA6843>