Date: Wed, 3 Jun 2009 22:58:10 +0200 (CEST) From: Wojciech Puchar <wojtek@wojtek.tensor.gdynia.pl> To: Reid Linnemann <lreid@webmail.cs.okstate.edu> Cc: freebsd-questions <freebsd-questions@freebsd.org> Subject: Re: An adage for gmirror users Message-ID: <alpine.BSF.2.00.0906032256170.27604@wojtek.tensor.gdynia.pl> In-Reply-To: <4A26DBAD.6000005@cs.okstate.edu> References: <4A26DBAD.6000005@cs.okstate.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
> My mirror gm0 consists of two SATA disks, ad4 and ad6. Now, I have a > finicky controller that sporadically spits out READ_DMA and READ_DMA48 or bad cables. > timeouts inexplicably. So at some point in time immemorial after > installing the last kernel, ad4 suffered a number of READ_DMA48 errors > and dropped out without being removed from the mirror's provider list . > So when I installed my new flashy kernel with all my filesystems > mounted, it was put into /boot/kernel on the mirror, which at that point > consisted of only ad6. On boot, the loader grabbed the kernel from i simply have in crontab a script running once per hour: #!/bin/sh /sbin/gmirror status|grep -q DEGRADED && \\ mail -s "gmirror failure" myphonenumber@mygsmoperator.pl </dev/null anyway - what a sense of using gmirror without regularly checking of failures at all?
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?alpine.BSF.2.00.0906032256170.27604>