Home
| Calendar
| Mail Lists
| List Archives
| Desktop SIG
| Hardware Hacking SIG
Wiki | Flickr | PicasaWeb | Video | Maps & Directions | Installfests | Keysignings Linux Cafe | Meeting Notes | Linux Links | Bling | About BLU |
You've shown that Dell software behaves differently on CentOS vs RHEL. You have *NOT* shown any evidence that Red Hat did any proprietary development for Dell. Jarod works at Red Hat and is in a postion to be aware of any proprietary development at Red Hat, if any existed. He has repeatedly asserted that no such development exists. On Fri, Dec 10, 2010 at 12:13 AM, Edward Ned Harvey <blu-Z8efaSeK1ezqlBn2x/YWAg at public.gmane.org> wrote: >> From: Jarod Wilson [mailto:jarod-ajLrJawYSntWk0Htik3J/w at public.gmane.org] >> >> >> On Dec 2, 2010, at 10:17 PM, Edward Ned Harvey wrote: >> >>> ... RH does proprietary development for Dell etc, regardless of what >> >>> anyone else might tell you. >> >> >> >> Thank you, I needed a good laugh. >> > >> > Not sure what that's supposed to mean, >> > but if you don't believe me, >> >> Its not that I don't believe you, its that you're WRONG. :) > > I've put a little bit of work into this, and here's what I found: > > In my vaguely recollected anecdote of a Dell support person identifying my > centos box by virtue of some closed-source library unavailable in centos... > At the time 7 yrs ago it was centos3/rhel3. ?So whatever the factual basis > in that memory, it's impossible to reproduce now, using hardware that > doesn't even remotely support an OS so old. ?Not like you'd care about > centos3/rhel3 anyway. > > In centos3, there are essentially no records of what's different between > centos3 and rhel3. ?But in more recent versions, 4 and 5, centos has done a > much better job of documenting the precise differences between centos and > redhat. ?If you just look in the release notes for centos4 or 5, they give > you a list of all the packages which they've modified, new packages they've > added, and rhel packages they've deleted. ?They say "All changes are > trademark and look/feel related unless otherwise noted under the specific > RPM." ... They list several dozen packages that have been changed, of which, > (centos 4.8 x86_64) 10 packages have substantial changes, and 8 packages > have been added, and none deleted. > > All of the 18 packages in question of centos 4.8 seem very harmless. ?For > example, changing the path from "/mnt/cdrom/RedHat" to "/mnt/cdrom/Centos" > and adding the SSL Cert Authority which certifies centos packages, and > replacing up2date via yum, and a few other things. ?It's pretty hard to > imagine anything running under rhel, being unable to run under centos. > > That being said, I created a multi-boot hard drive on dell hardware today, > on a system ?which I need to update the firmware. ?I ran the firmware update > and OMSA installer on centos 4.8, rhel 4.8, centos 5.5, and rhel 5.5. > Naturally, all the RHEL versions work flawlessly, but the question is > centos... ?And the answer is "sometimes." > > I don't know if Dell is employing their own software engineers, or if redhat > is doing development work for them, or any variation in between. ?But as far > as I can tell, the source code for these dell installation packages is > closed source, so it's very difficult to pinpoint the root cause of centos > failure. ?In some cases, I just edit an installation script, find the place > where it detects the RHEL os by /etc/redhat-release, and modify the script > to think it's the related version of RHEL... and problem solved. ?In other > cases, I literally have an ELF binary, which behaves differently for no > apparent reason, if I run it on centos versus redhat. > > One more thing: ?On the centos4 installation, I simply overwrote the > /etc/redhat-release file using the one from rhel4. ?And then the binary bios > detection executable worked. ?So I can only conclude that the binary has > hard-coded a check for the /etc/redhat-release file, just as some other > scripts have done. > > Regardless of centos vs redhat, my system is normally a solaris system, and > solaris is a "supported" os on this hardware ... yet solaris has no option > to apply firmware updates etc... ?All I needed to do was call Dell, and they > sent me an ISO for a Dell customized Centos LiveCD, which fully supports all > the firmware update packages and diagnostic tools. > > I take this to mean, you can safely assume you're able to apply firmware > updates, via livecd if necessary. ?And that basically leaves just OMSA. > Well, the OMSA installer was the one where I only needed to modify the OS > detection if-clause. ?It's a trivial modification, but it needs to be done > for both centos 4 and 5. ?The OMSA installer fails natively, on both centos > 4 and 5. > > So in conclusion: > > If you are going to use centos, I recommend either (a) overwriting the > /etc/redhat-release file using the supposed equivalent from RHEL, or (b) > you'll have to modify the OMSA installation script by hand, to accept centos > instead of redhat. > > If you opt for centos instead of rhel on dell hardware, it is NOT a safe > assumption that you can run OMSA, or firmware update packages natively. ?But > some things ARE safe to assume: ?(1) You can safely assume you're able to > apply all firmware updates. ?Via live CD if necessary. ?In fact, DON'T > attempt firmware updates any way other than live CD. ?I had a firmware scare > today, centos 4 crashing in the middle of firmware update after I "tricked" > it into thinking it was RHEL4. ?Fortunately the system was still bootable, > with signs of instability, and I just flashed it again using genuine RHEL4. > Everything smooth and back to normal. ?(2) ?I only tested one system, but in > my opinion, you can probably safely assume OMSA and requisite drivers will > run on centos 5, with perhaps a little bit of manual intervention such as > overwriting /etc/redhat-release. ?I am less confident about centos 4, if you > care. > > It is not clear the reasons why. ?In fact, if any of you want, I can give > you the "ldd" and "strace" outputs I obtained from the BIOS version > detection elf binary today. ?Suffice it to say: ?I examined it thoroughly, > and all I was really able to conclude was that the > bios-version-detection-binary "biosie.bin" (and some other binaries) behave > dramatically different on supposedly identical versions of centos versus > rhel. ?It works flawlessly on rhel 4.8, and it fails entirely on centos 4.8. > The end result is inability to update firmware on centos 4.8. ?I am required > to use either Dell customized LiveCD, or some other means to apply the > firmware update on centos 4.8. > > > _______________________________________________ > Discuss mailing list > Discuss-mNDKBlG2WHs at public.gmane.org > http://lists.blu.org/mailman/listinfo/discuss > -- John Abreau / Executive Director, Boston Linux & Unix AIM abreauj / JABBER jabr-iMZfmuK6BGBxLiRVyXs8+g at public.gmane.org / YAHOO abreauj / SKYPE zusa_it_mgr Email jabr-mNDKBlG2WHs at public.gmane.org / WWW http://www.abreau.net / PGP-Key-ID 0xD5C7B5D9 PGP-Key-Fingerprint 72 FB 39 4F 3C 3B D6 5B E0 C8 5A 6E F1 2C BE 99
BLU is a member of BostonUserGroups | |
We also thank MIT for the use of their facilities. |