Re: [Grml] BadRam/BadMem Kernelpatch
So IMO this would only be useful for private use.
i don`t think so - at least i wouldn`t say it`s only usefule for private use if you mean "home use"
- think of developing countries where people cannot afford replacing their ram because of just "one bit out of millions" is bad
- think of companies with large test/development environments which are not mission critical. we have many boxes out of service contract at work - and we still use them until they die or getting too old
- think of a server which constantly crashes during a recover you try with grml - due to bad ram. should you really buy new ram for that machine if it gets replaced, anyway - and you just need it being running stable for the recovery process ? maybe this is just a theoretical scenario, but the discussion about the pro`s and con`s of badram patch is very controversial.
anyway - i heard there will be an request for inclusion into mainline soon.
roland
-----Ursprüngliche Nachricht----- Von: Andreas Gredler jimmy@grml.org Gesendet: 30.10.06 13:55:46 An: grml@mur.at Betreff: Re: [Grml] BadRam/BadMem Kernelpatch
On Sun, Oct 29, 2006 at 03:02:15PM +0100, devzero@web.de wrote:
Description:
BadRAM: Linux kernel support for broken RAM modules Summary: This page proposes an approach to support RAMs with defective addresses, This may open interesting business perspectives, where those RAMs can be sold under a white label for less money rather than discarded of without any profit.
If you talk about business... Broken RAM always gets replaced by new modules, since this is mission critical and covered by warranty and/or service contracts. Additionally, "real" servers mirror RAM and have spare RAMs installed. So IMO this would only be useful for private use.
greets Jimmy
-- Andreas "Jimmy" Gredler ,'"`. http://www.jimmy.co.at/ | jimmy@g-tec.co.at ( grml.org -» Linux Live-CD for texttool-users and sysadmins `._, http://www.grml.org/ | jimmy@grml.org _______________________________________________ Grml mailing list - Grml@mur.at http://lists.mur.at/mailman/listinfo/grml join #grml on irc.freenode.org grml-devel-blog: http://grml.supersized.org/
_____________________________________________________________________ Der WEB.DE SmartSurfer hilft bis zu 70% Ihrer Onlinekosten zu sparen! http://smartsurfer.web.de/?mc=100071&distributionid=000000000066
On Tue, Oct 31, 2006 at 02:04:50PM +0100, devzero@web.de wrote:
So IMO this would only be useful for private use.
i don`t think so - at least i wouldn`t say it`s only usefule for private use if you mean "home use"
- think of developing countries where people cannot afford replacing
their ram because of just "one bit out of millions" is bad
I don't know about service contracts and warranties in such countries. So maybe you're right.
- think of companies with large test/development environments which
are not mission critical. we have many boxes out of service contract at work - and we still use them until they die or getting too old
This is a rare business case IMO, because when the boxes are out of service contracts they get replaced or disposed, usually. But I've seen such boxes, too ;-)
- think of a server which constantly crashes during a recover you try
with grml - due to bad ram. should you really buy new ram for that machine if it gets replaced, anyway - and you just need it being running stable for the recovery process ? maybe this is just a theoretical scenario, but the discussion about the pro`s and con`s of badram patch is very controversial.
Most times it's sufficient to remove one RAM module and to continue with less RAM to recover. But you're right. This seems to be a good case, where it might be important for business, too.
Maybe I should rewrite "So IMO this would only be useful for private use" to "So IMO this would only be useful for private use and small business"
Anyway, if it's planned for mainline we should wait for it.
greets Jimmy
Teilnehmer (2)
-
Andreas Gredler -
devzero＠web.de