Re: Spin-up maxtor onetouch

--- In nslu2-general%40yahoogroups.com">nslu2-generalyahoogroups.com, "David Shepherd&quot; <david...> wrote:
&gt;
> I'm using Unslung 6.10 and after following instructions on how to move
>; /var and /dev to a ramdisk + added noatime to the disk mounts I
> managed to get my Maxtor Onetouch III to spin down (no 5 sec flicker
> on disk light + after 10-15 mins disk definitely span down)! However,
> when I looked at the slug a couple of days later found it no longer
&gt; thought it had a USB disk attached + attempts to login into it or
> access it via web interface failed.
>
> Looking at nslu2-linux.org saw comments abut IO requests timing out
> when Seagate freeagents had spun down along with a "fix&quot; involving
> creating an auto_restatr file in a dir under /sys/ ... however, this
>; seems to be for Debian and I'm on Unslung ... any ideas whether
> there's an equivalent fix I can use.
>;
> If not ... while it seems "tidy" to be able to get the Onetouch to
> spindown (its mainly aimed at being a backup server which will
>; eventually run rsnapshot once or twice a day to snapshot my main
>; active NAS drive slo its not really in use much at the moment) is it
> really necessary to get the drive to spin down .... will it affect
&gt; reliability much etc?

Update on how I got on.

After getting the disk to spin down and the finding it wouldn't spin
up I restarted everything and didin't do the copy /var to ram disk
trick again ... however I found that after a day or so the NSLU2 was
crashed with a spun down disk. I tried reformatting disk and
re-unslinging etc but same thing happens,

Eventually I took the "last resort&quot; and used the LinkSys Eraseall tool
to revert to (as close as possible to) its initial state. This worked
fine.

N.b. as is the habit of things on internet the whole EraseAll
situation has become totally exagerated and when googling to find out
how to use it I found endless websites which solemnly declare that if
you use eraseall then your NSLU2 will be permanently bricked! Given
that its a tool produced by Linksys this would seem unlikely but
common sense is not a major asset in parts of the web! Clearly
eraseall is more dangerous than other toosl to reflash the NSLU2 since
it apparent rewrites all the redboot code so if something does go
wrong you may have bricked the NSLU2 so for standard unslinging etc
its safer to use something less drastic ... however, in my situation
(to quote a UK TV advert) "it did what it says on the can".

With the NSLU2 back to standard Linksys stae I was able to unsling and
so far after a week or so its all working ok. Maxtor disk is not
spinning down but I'm not seeing that as a problem ... in fact if
anything when lookinhg for info on any issues from 24/7 running on the
Seagate/Maxtor web site the info given there was definitely that disk
are designed to be able to run constantly and if there is an issue its
with constant spin down/spin up cycles!

Again I think I'd fallen for "web folklore&quot; as I recall some of the
original IBM deskstars had reliabiliyt problems which IBM at the time
tried to explain by saying "well, they're desktop drives which aren't
meant to be used more than 8 hours a day".

Messages in this topic (7) Reply (via web post) | Start a new topic

.

David Shepherd wrote:
&gt; N.b. as is the habit of things on internet the whole EraseAll
> situation has become totally exagerated and when googling to find out
> how to use it I found endless websites which solemnly declare that if
> you use eraseall then your NSLU2 will be permanently bricked! Given
&gt; that its a tool produced by Linksys this would seem unlikely but
> common sense is not a major asset in parts of the web!

It is a fact that the *only* reported cases of end-user brickage of
NSLU2 units have occurred when the EraseAll tool has been used. There
have been no confirmed reported cases of permanent (i.e. requiring
hardware JTAG to fix) brickage when people use UpSlug2 or the SerComm
updater to reflash as recommended by the nslu2-linux project.

There are at least two cases where EraseAll is guaranteed to brick your
device:

1) You get a power failure during flashing. Note that power failure
includes the cat knocking the power adapter out of the wall.

2) You try and flash a Debian image (which for DFSG-compliance reasons
does not contain a replacement copy of RedBoot).

> Clearly eraseall is more dangerous than other toosl to reflash the NSLU2 since
&gt; it apparent rewrites all the redboot code so if something does go
> wrong you may have bricked the NSLU2 so for standard unslinging etc
> its safer to use something less drastic ... however, in my situation
> (to quote a UK TV advert) "it did what it says on the can".

A goal of the nslu2-linux project is to not increase the warranty/RMA
return rate of NSLU2 devices to Linksys. At the moment, the only cases
we know where using nslu2-linux firmware has caused a return is when
someone has used the EraseAll tool and encountered one of the two known
failure modes.

For this reason, we strongly advise, at every opportunity possible, that
people do *not* use EraseAll.

Now, anyone is free to disregard that advice if they know what they are
doing. At that point, any failure becomes the sole resposibility of the
person making that choice - the project has done all it can to prevent
problems. We don't prevent people using EraseAll, and if it does the
job for someone who fully understands the risks, then that's great.

-- Rod Whitby
-- NSLU2-Linux Project Lead

Messages in this topic (8) Reply (via web post) | Start a new topic

.