On May 11, 2010, at 17:18, "Rod Whitworth" <glis...@witworx.com> wrote:
> On Tue, 11 May 2010 12:43:17 -0400, Chris Smith > > I have tried to kill a CF for years. For more than a year it was > running spamd with the most verbose logging possible and lots of other > read/writes the system could live without. > > It is still going. > > I suggest that you use CF and when upgrade time comes around you > program a new one and then have a halt-swap-reboot event and send me > the one you don't think has much life left. I'll try wearing it out for > you. > > My clients have lost more hard drives last year (3) than CFs in my > lifetime (0) and I've been using them since they were exorbitantly > priced. > > Some of that is good luck but they sure are not easily worn out. I'd have to agree there. I had one CF fail after three years of heavy DNS logging and I had a brand new card fail immediately as well. I've had many more times the hard drives fail. I would also suggest looking at the flashrd project. http://www.nmedia.net/flashrd/ I just recently started using it on some individual firewalls as well as several clusters. The whole point of the setup is to mount everything possible as read only and the rest to mfs. Bryan
