@pfo IBM. Wasted half a day trying to figure out why my dnsmasq PXE setup was just not working and dug deep into it this morning. Found out it's not sending the architecture I was expecting :( sad panda
I know Intel makes a compiler for it, but not sure if it's an open spec or where to find it...
@Iain Mediocre. I didn't have enough material to cover for the first class, so they got out an hour early.
I was pretty nervous starting the class but settled in after the first hour. It should be much better over the next few weeks as I get into the swing of things.
@ChrisS sourceforge.net/apps/mediawiki/tianocore/index.php?title=EDK2 or gnu-efi seems to be able to build EFI apps. Not sure if it's compiling to EFI BC or EFI x64 though. I think I've been sending back EFI x64 binaries to machines that have been requesting EFI BC, but I guess they're just doing the Right Thing. Or I'm actually compiling to EBC.
It was supposed to be a 4 hour class. The previous teacher spent the first night only on chapter one (and intro to class stuff). I did all that and Chapter 2. And still ran out of material... Not sure WTF he was doing for that 4 hours.
@MikeyB I would think the machines would fall back to executing x64 code if it isn't EBC, but I'd have any actual experience in what you're doing.
... though it's interesting as hell and you'd better keep us updated with what you're figuring out.
This seemed to have happened after a server restart and a server restart seems to temporarily solve the problem. Everyone's database does not crash at the same time. Mine may be working while someone elses is down. When it is down a test run from SQL in ODBC fails. I am not very familiar with ...
apparently symcbean didn't like my simplified answer to the guy trying to use mail@mydomain.com instead of just his username.
...no duh you can alias mail through whatever mechanism your MTA uses. Derp. But this guy was obviously new, not knowledgable, and I wouldn't want to answer his next questions on errors and login issues when he'll fail to distinguish the mail user from the mail alias he's cobbled together on his system.
Instead he downvoted mine called my simplified explanation to a comment nonsense and gives his own answer to encourage him to make it more complicated. peh
/kicks chair - sometimes just doesn't pay to get out of bed. Lousy mood.
So, I've got it working, but on the downside it looks like the code that constructs the PXE boot menu on IBM UEFI systems is absolute shite. Or I'm doing it wrong.
My position was to just steer him away from it as a bad idea, since he very well might run into problems with troubleshooting and logging and not make distinctions among the mail user, mail components, mail alias, etc. and it sounded like his reason for using mail@... might have been kind of silly to begin with. Just use your username and forget about it.
Really I think your wit and humor take some getting used to. You also seem to be a little less on the...what's the right term? You don't couch opinions in niceties as much, in general.
We do something quite similar here. 3 subnets behind a CentOS5 "router". Basically we just have iptables set to the follow 'nat' table rule:
iptables -t nat -A POSTROUTING -o -j SNAT --to-source
In our case, device is eth1 and the IP is 10.0.0.2 to differentiate from the Class C IP4 subnets w...
@MDMarra In seriousness, at the very least there is now someone familiar with this community that is in this chat a lot to handle flags before we get another chatpocalpyse
@pfo The menu appears over top of text already drawn and you can't select anything. OTOH, I think something's going wrong in my config or dnsmasq as no bootfile options (pxe opt 71) are being sent. I think this may be confusing the PXE client.
@MikeyB thanks, that would be great. I wonder that you got elilo to work since it's basically an abandoned project - it was started before IBM was selling uEFI machinery.
@pfo Looks like it hasn't been updated (releases that is) in a year but that's not too bad. And until efi-grub supports loading things over the network. I had no trouble getting elilo to work (other than the broken documentation re: the .msg file)
@MikeyB I've seen a nice UI and http library for EFI a few weeks ago and was thinking about trying to get something running on my own and I'll play around next weekend to see how that is going
@LucasKauffman EFI applications can come from a variety of sources; adapter cards (what used to be a "BIOS Option ROM" is now an EFI Application (EFI Driver, specifically) that gets loaded from the card), adapter card configuration utilities, network (DHCP), onboard flash, hard disks (usually bootloaders)
EFI (technically *U*EFI) is effectively just a BIOS replacement with some trimmings -- The only real "useful app" on top of it is a boot loader. One that supports Forth so you can play BootPong :P
@LucasKauffman Being able to load an EFI Application to configure say, your RAID card, and then exit back into the EFI Setup without rebooting is really sexy. And that EFI Application comes from the RAID card itself.
@voretaq7 Well IBM for sure, presumably Dell and HP boxen do it the same way. If you're still waiting until the 'Press Ctrl-C to configure this card' you're doing it wrong.
@LucasKauffman . . . It's not written in Forth, it doesn't count :P
(also: Seriously? SERIOUSLY??? Who thought that would be a good thing to include in a boot loader?!?... not stopping me from playing it mind you, but still)
@WesleyDavid I think it's a great marketing campaign. "If you become a Network Admin you will be surrounded by attractive women who are into light bondage."
Seriously, Novell Support? "I can't access EFI boot variables after booting up SLES11 via EFI over the network." "Here's how to configure network booting with EFI. On a Dell." "YOU'RE SOLVING THE WRONG PROBLEM!" *facepalm*
@voretaq7 Makes my users crazy. Takes 7 seconds to shutdown services and 12 to start Ubuntu on that machine, but most of the TTR is the BIOS controller moseying through its boot process.
Proposed Q&A site for people who are enthusiastic about fashion, and about clothing, tailoring, cobbling, hatmaking, haberdashery and bespoke or couture clothing items.
@voretaq7 Yeah. We used to run our internal web app on an old dual-core with 2.5GB of RAM. It was painful. Everyone was amazed at how fast it ran with 24GB. RAM for Read caching FTW.
(I wouldn't say I'm "enthusiastic" about fashion, but I'd like to be able to sanity check and make sure I'm not a total fashion disaster when I walk out the door :P)
@voretaq7 They don't want to have to wait for me and my predecessor was a) incredibly lazy and b) an unmitigated asshole. I'm considered more of a mitigated PITA, as near as I can tell.
I think we ought to have an SE site about coffee. Proper milk temperatures for lattes, how to pour the milk properly, correct pastry pairing with various drinks, proper care of espresso equipment and french presses.
Hey everyone. I need to renew our Wildcard SSL cert. We're currently with RapidSSL (GeoTruet). If anyone has any recommendations for other providers I'd be glad to hear it before I renew. Thanks!
@voretaq7 Thats how I almost bought a SPARC once with 32 cores and 32 GB of RAM, was like 500 euro at the time. It was huge though and consumed shitloads of power :p
@TylerShads You're a blue. Can you tell me whether it would be a violation of the .SE terms of service to hold a raffle to select a person to go beat Shuttleworth with a shovel?
@TylerShads I want to hold a raffle to select someone who's local to the UK to go beat Mark Shuttleworth with a shovel. One with a long narrow blade that is flat at the mouth and round along its lengthwise axis.
@TylerShads I'm sure Ububtu did something to offend @AdrianK. It's like the Joe Biden of operating systems: it can't go 5 minutes without offending someone.
@ScottPack there's too many puppeteers on SF and SO now. I subscribe to puppet and now every time there's a solid answer before I get a chance to jump in and write one. :)
(at least for the easy, well-written, clear questions)
Last night I started writing up an answer, saved a first draft, then while fleshing it out I realized that I had misread the question, so my answer was entirely irrelevant. So I deleted it.
@pfo Not necessarily. I think most puppet questions would fare better on SF, but there's some that are really about the programming and are on-topic on SO and there's no need to migrate those.
If I'd had questions about writing custom facts or custom functions, I'd ask it on SO, since it's really a ruby in a special library environment thing and not a managing servers thing
@TylerShads It's actually a trick we played on my first boss back when I was a consultant. He was in the datacenter for something and we played a blood-curdling scream through one of the Sun servers in the rack behind him.
If I understand your question correctly your situation is that the old hosting provider will not give you the SSL certificate and private key you are currently using? That seems VERY shady to me.
Having worked at an ISP/MSP that did hosting and SSL Certificate registration I can see no legitimat...
(Honestly when I would give people their keys at the ISP I would advise them to call Verisign and have the cert revoked/reissued with a new key anyway because even though they could trust us best practices say they shouldn't)
wait... when stuff exits it tells me what happened? MADNESS!
There is an excellent website for this here
RFC 4578 defines various machine architectures for PXE as follows:
Type Architecture Name
---- -----------------
0 Intel x86PC
1 NEC/PC98
2 EFI Itanium
3 DEC Alpha
4 Arc x86
5 Intel Lean Client
6 EFI IA32
7 EFI BC
8 EFI Xscale
9 EFI x86-64
...
This seemed to have happened after a server restart and a server restart seems to temporarily solve the problem. Everyone's database does not crash at the same time. Mine may be working while someone elses is down. When it is down a test run from SQL in ODBC fails. I am not very familiar with ...
We do something quite similar here. 3 subnets behind a CentOS5 "router". Basically we just have iptables set to the follow 'nat' table rule:
iptables -t nat -A POSTROUTING -o -j SNAT --to-source
In our case, device is eth1 and the IP is 10.0.0.2 to differentiate from the Class C IP4 subnets w...
