[j-nsp] High failure rates for M7i/M10i hard disks?

[Richard A Steenbergen]

On Mon, Aug 22, 2005 at 08:21:35AM +0100, Jee Kay wrote:
> On 8/22/05, Johannes Resch <jr at xor.at> wrote:
> 
> > to add another datapoint: within the last year, we had to RMA two RE-2.0
> > because of failure of internal flash; opposed to zero HD-failures.
> 
> We just bought 5 new M7is.... two months later, 3 have had their REs
> RMAd with HDD failures. Needless to say we're now refusing to put
> these shoddy pieces of equipment into production until we get some
> sort of explanation extracted from Juniper.
> 
> This together with the absolutely awful 'product' called the J6300
> (bought 3, all 3 had hardware problems and constant crashes) is
> seriously turning me off Juniper.
> http://puck.nether.net/mailman/listinfo/juniper-nsp

What model hard drive are they using on the M7i RE-5.0 routing engines 
these days? AFAIK the M#i model RE's (5.0/5.0+) were custom-made by 
Kontron explicitly for Juniper, and aren't just existing products. I 
personally can't say enough bad things about the old IBM Travelstars they 
used to use, I have quite the stack of dead ones.

Not to say that hard drives shouldn't be in routers, personally I think it 
adds an amazing amount of functionality which is worth a little bit of 
extra risk. I've probably replaced both the hard drives and compact flash 
from around 20 RE-2.0s and 3.0s, out of hundreds of them that I have run 
across, including some really roughly handled gear that had been abused in 
shipping and probably in its prior life. In my experience the CF fails 
more often than the hard-drive just through age, even with Travelstars 
involved.

Really sounds like there may be a production problem with the RE-5.0's,

-- 
Richard A Steenbergen <ras at e-gerbil.net>       http://www.e-gerbil.net/ras
GPG Key ID: 0xF8B12CBC (7535 7F59 8204 ED1F CC1C 53AF 4C41 5ECA F8B1 2CBC)