Re: [SLE] quick shutdown
not they're in use) and have ACPI or APM or whatever switch
the machine off.
Is this possible? Any pointers appreciated /sbin/halt.
This still does the stopping_this_service stopping_that_service and takes a good 30 seconds before the machine switches off. This is too long. I want it to get the "switch-off-now!" signal, unmount all filesystems regardless of weather they're being used or not, and cut the power. Or better - cut the power straight away. This is emergency procedure. My data is important but NOT as valuable as the hardware. Thanks Hans == Download ringtones, logos and picture messages at Ananzi Mobile Fun. http://www.ananzi.co.za/cgi-bin/goto.pl?mobile
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Sunday 16 March 2003 11:43 am, H du Plooy wrote:
This still does the stopping_this_service stopping_that_service and takes a good 30 seconds before the machine switches off. This is too long.
I want it to get the "switch-off-now!" signal, unmount all filesystems regardless of weather they're being used or not, and cut the power. Or better - cut the power straight away. This is emergency procedure. My data is important but NOT as valuable as the hardware.
I think you've just given yourself the answer: use the switch on the wall [powerstrip] to cut power to the system and pick up the pieces later. Offhand I can't think of many reasons why you would want this -- some fanciful ones come to mind [the CPU is controlling a robotic arm that is about to crush the system itself...] but that's about it -- you state this is an emergency procedure [the building is on fire?] and that the hardware is more important than the data [very rare -- implies you have religiously applied a backup & recovery system] but most "emergency shutdown" scenarios I can think of (like a fire) imply the external destruction of the hardware anyway... However, just before I hit "send" I thought of another avenue you might take: look into the "powerfail" shutdown mode/script/whatever [something like runlevel 6, I think?] This is supposed to be triggered by a signal from an battery-backed power supply that (a) power is out, and (b) the battery is nearly depleted... The implication being that there may not be enough power to flush disk write queues if the system keeps "processing" for much longer... - -- Yet another Blog: http://osnut.homelinux.net -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.7 (GNU/Linux) iD8DBQE+dN8iV/YHUqq2SwsRAh5lAKC+mR6XukwvbUOMNzuwBLKA0yUBEwCbBYW4 nfMUdiTJYTdTqDZxMU/SNMA= =HFUD -----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Sunday 16 March 2003 12:31 pm, Tom Emerson wrote:
On Sunday 16 March 2003 11:43 am, H du Plooy wrote:
This still does the stopping_this_service stopping_that_service and takes a good 30 seconds before the machine switches off. This is too long.
thought of one more [but haven't tested it] from a console [i.e., outside of "X"], press ctl-alt-delete -- this should flush queues and jump to the "reboot" routine in your BIOS, and while the system is going through the "POST" operation, you can safely hit the power switch to shut the system off - -- that should cut your shutdown time to roughly 3-5 seconds - -- Yet another Blog: http://osnut.homelinux.net -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.7 (GNU/Linux) iD8DBQE+dOWSV/YHUqq2SwsRApcyAJ9WwMqwMefHGPZQlgG9TXyBjS6+hACdG3MS W1bNFkkPr6uJUWxbrCo+y7Q= =mn3e -----END PGP SIGNATURE-----
thought of one more [but haven't tested it] from a console [i.e., outside of "X"], press ctl-alt-delete -- this should flush queues and jump to the "reboot" routine in your BIOS, and while the system is going through the "POST" operation, you can safely hit the power switch to shut the system off - -- that should cut your shutdown time to roughly 3-5 seconds
If memory serves, all recent distros map C-A-D to /sbin/reboot or the equiv. No more of that "cowboy reboot" stuff allowed like in DOS. -- John LeMay KC2KTH Senior Enterprise Consultant NJMC | http://www.njmc.com | Phone 732-557-4848 Specializing in Microsoft and Unix based solutions
I want it to get the "switch-off-now!" signal, unmount all filesystems regardless of weather they're being used or not, and cut the power. Or better - cut the power straight away. This is emergency procedure. My data is important but NOT as valuable as the hardware.
It may help to explain why you are trying to do this. Looks like I missed the original post since I don't see it in my inbox. In an emergency situation, halting the system is - as you know - the best solution for data protection and when the machine is connected to an adequate UPS power is not an issue. Fire or other natural disaster shouldn't be an issue here - if you are running to save your life I hope you aren't stopping in the computer room on the way out! Your opinion that your hardware is more valuable than your data is quite different than any other opinion I've heard. I suppose it begs the question - why store the data anyplace if it isn't that valuable? Hardware can always be had relatively cheaply, recreating data isn't free unless your time has no value (to coin an old phrase). -- John LeMay KC2KTH Senior Enterprise Consultant NJMC | http://www.njmc.com | Phone 732-557-4848 Specializing in Microsoft and Unix based solutions
On Sunday 16 March 2003 03:55 pm, John LeMay wrote:
I want it to get the "switch-off-now!" signal, unmount all filesystems regardless of weather they're being used or not, and cut the power. Or better - cut the power straight away. This is emergency procedure. My data is important but NOT as valuable as the hardware.
It may help to explain why you are trying to do this. Looks like I missed the original post since I don't see it in my inbox. In an emergency situation, halting the system is - as you know - the best solution for data protection and when the machine is connected to an adequate UPS power is not an issue. Fire or other natural disaster shouldn't be an issue here - if you are running to save your life I hope you aren't stopping in the computer room on the way out!
Your opinion that your hardware is more valuable than your data is quite different than any other opinion I've heard. I suppose it begs the question - why store the data anyplace if it isn't that valuable?
I have a similar situation. I do web programming. All of my data, other than my emails and some local personal files, are stored on the servers I work on. I drag my files into my editor from my ftp client and then save them directly back to the server. I don't store anything locally because I don't have UPS and I often mess around with my systems to the point that I have to reinstall. It's much easier if I don't have important data on my machine that I may loose. My hardware on the other hand is expensive and if my processors were overheating (which could happen since I run dual AMD processors in each box), I would want the computer to be shut down immediately. Any data that I lose is going to be easy to replace. Having to buy new processors and/or other hardware would not be so easy for me to replace. -Gary
Hardware can always be had relatively cheaply, recreating data isn't free unless your time has no value (to coin an old phrase).
-- John LeMay KC2KTH Senior Enterprise Consultant NJMC | http://www.njmc.com | Phone 732-557-4848 Specializing in Microsoft and Unix based solutions
Your opinion that your hardware is more valuable than your data is quite different than any other opinion I've heard. I suppose it begs the question - why store the data anyplace if it isn't that valuable?
See earlier post on this. I keep documents I make (which I keep incase I need to refer to them later, but in the past four years that has happened maybe twice), on removeable disk in duplicate. I have most of my CDs on my computer in mp3/ogg. That can be made again. e-mail: backed up weekly, not too serious either. My point is, in my situation, my hardware is worth more to me than the stuff on my hard disk. Hans == Download ringtones, logos and picture messages at Ananzi Mobile Fun. http://www.ananzi.co.za/cgi-bin/goto.pl?mobile
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Sunday 16 March 2003 12:55 pm, John LeMay wrote: [in response to H du Plooy]
I want it to get the "switch-off-now!" signal, [...]
It may help to explain why you are trying to do this. Looks like I missed the original post since I don't see it in my inbox. In an emergency situation, halting the system is - as you know - the best solution for data protection
I went back a couple of threads and checked -- seems he did give a reason: On Tuesday 11 March 2003 8:06 am, H du Plooy wrote:
I'm trying to get a script going that would shutdown my linux system in case of emergency (like the CPU fan going bust).
I think if the CPU fan is going/goes, the system may shut itself down one way or another [however, unless you have an alarm on the CPU fan, most likely the only way you're going to know it stopped is when you smell/see smoke...] - -- Yet another Blog: http://osnut.homelinux.net -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.7 (GNU/Linux) iD8DBQE+dOqAV/YHUqq2SwsRAuknAJ0VshOJw183EgaqzSmk9pMJdM1MtQCeIcVG nJucrdd+27DfjBgoAYI7WP4= =o6oc -----END PGP SIGNATURE-----
I'm trying to get a script going that would shutdown my
linux system in case of emergency (like the CPU fan going bust).
I think if the CPU fan is going/goes, the system may shut itself down one way or another [however, unless you have an alarm on the CPU fan, most likely the only way you're going to know it stopped is when you smell/see smoke...]
Thanks Tom. Actually, Intel processors will shut themselves down in the case of a thermal emergency like this. I've had it happen twice - once on a Celeron 400 and once on a P3-1GHz. I don't think AMD has implemented this yet. -- John LeMay KC2KTH Senior Enterprise Consultant NJMC | http://www.njmc.com | Phone 732-557-4848 Specializing in Microsoft and Unix based solutions
On Sunday 16 March 2003 22:25, John LeMay wrote:
Thanks Tom. Actually, Intel processors will shut themselves down in the case of a thermal emergency like this. I've had it happen twice - once on a Celeron 400 and once on a P3-1GHz. I don't think AMD has implemented this yet.
It's been in AMD based systems for a long while, although it didn't work so well in the past. When Tom's hardware tested it on a 1GHz Thunderbird, the cpu physically melted before the motherboard had a chance to power off. Rumour has it that it works better with newer athlons, but I've been fortunate enough not to have to try it :)
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
It's been in AMD based systems for a long while, although it didn't work so well in the past. When Tom's hardware tested it on a 1GHz Thunderbird, the cpu physically melted before the motherboard had a chance to power off. Rumour has it that it works better with newer athlons, but I've been fortunate enough not to have to try it :)
Yes, the AMDs are so thermally frigile that on average if something like the cpu fan seating is done improperly the the cpu will literally fry in under 3 seconds. Most of the time this particular situation happens at boot up and by the time the system boots to the point that it can give a warning (a POST) the owner will smell something burning and in some cases I had friends tell me that they immediately looked over at the case to see smoke coming from it. I had the guy at the shop where I bought my board and cpu install the cpu and fan to the board - he said that if it fried or similar that he could cover it, but if I put it in that I was on my own. Presently my system has 3 system fans (one is the same size as a cdrom - can't remember the dimensions), one cpu fan (of course) and 2 hard drive fans. It's still runs a tad on the warm side. I really like the AMD cpu - just wish they weren't so fragile. The upside to pentiums/celerons is that they keep chugging away. I had a cpu fan become partially detached - I noticed it when my system became really slow and started to give mem errors. I turned of the systems for a bit, reattached the fan, and didn't have any further issues. Cheers, Curtis. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.7 (GNU/Linux) iD8DBQE+dPsV7WVLiDrqeksRAhkvAJ0aGe9HlcRWCEKsVPdr7qv1SrbrDwCgmoDf 5YYH5zPY4lA+BDThZKuSjjs= =clbp -----END PGP SIGNATURE-----
Rumour has it that it works better with newer athlons, but I've been fortunate enough not to have to try it
Exactly. I recall the Tom's Hardware article you mentioned. When I saw the Intel CPU's shut themselves down upon a fan failure I was sold on Intel. Not that I hadn't been for a while though. I haven't used a non-Intel CPU since the Cyrix 200+, the awful beast it was. I'm so sold on Intel that I even switched to thier system boards (I know, Intel doesn't make them) after my experience with Abit boards with failing capacitors. Intel may not make the board, but when one fails I have a new one the next day. Abit takes months. -- John LeMay KC2KTH Senior Enterprise Consultant NJMC | http://www.njmc.com | Phone 732-557-4848 Specializing in Microsoft and Unix based solutions
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
after my experience with Abit boards with failing capacitors. Intel may not make the board, but when one fails I have a new one the next day. Abit takes months.
Sheesh, I thought I was one of the few that had Abit capacitor faliures. I had an BM-6 Celeron board that one day just wouldn't boot. The said most likely the bios failed. So, I bought a BE-6 II, rev 2.0. I had it for about 18 months and the same exact thing happened. I couldn't figure out what the hell was going on. But decided to give Gigabyte a try. as well as AMD XP cpu. When I bought the system (the one I have now) the guy said he used to love ABit. I thought they were really nice boards myself, friends of mine recommended them. He then enlightened me about ABit switching suppliers of some of their parts and that's when his customers started to have a lot of problems with failing boards - so he doesn't recommend them or use them personally anymore. Sure enough, when I got home I looked at both my boards and you could see where the tops of the capacitors actually "popped" at the top of two or three of them. Pity, I really like my ABit boards right to about the 1-1/2 year point - when both failed, and the difference between both the fails was about a week or two apart (e.g ~ +/- week within 18 months). Cheers, Curtis. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.7 (GNU/Linux) iD8DBQE+dP+87WVLiDrqeksRAiZ9AJ9cJGXqgWokWc1qOl/U7MJtPdEM9ACeNtIR 25v2joGNz763SsVPml6ldRo= =5emt -----END PGP SIGNATURE-----
No, you were one of many with bad caps on boards. There was a whole story about how it happened. An engineer or engineers from the plant that made components - ie capacitors - left and "stole" the formula for the product. Turns out they stole a faulty formula and manufactured bad parts. They unknowingly sold a lot of these for quite a while - lots of them went to Abit it would seem - before finding out about the issue. It's kept "Homie" in business for a while now: http://home.att.net/~garyheadlee/index.html I haven't used Gary - nor do I know him - but I've cooresponded with him about the issue a couple of times. -- John LeMay KC2KTH Senior Enterprise Consultant NJMC | http://www.njmc.com | Phone 732-557-4848 Specializing in Microsoft and Unix based solutions
On Sun, 16 Mar 2003 16:25:23 -0500
John LeMay
Thanks Tom. Actually, Intel processors will shut themselves down in the case of a thermal emergency like this. I've had it happen twice - once on a Celeron 400 and once on a P3-1GHz. I don't think AMD has implemented this yet.
AMD actually requires motherboard manufacturers to put circuitry on their boards that would do this (it's not built into the chip), but only a few boards by ASUS and I *think* Soltek actually do this. http://www6.tomshardware.com/column/20011029/index.html Hans == Download ringtones, logos and picture messages at Ananzi Mobile Fun. http://www.ananzi.co.za/cgi-bin/goto.pl?mobile
On Sun, 16 Mar 2003 15:55:57 -0500
John LeMay
It may help to explain why you are trying to do this. So that when I'm away from my machine, and the cpu fan decides to give the ghost, the machine will power down by itself, and I won't have to replace anything.
In an emergency situation, halting the system is - as you know - the best solution for data protection I'm not interested in data protection. The worst thing I could loose is the seti packet the machine was working on at the time. Anything important I always have a backup of on either stiffy or CD-RW.
I can understand how large corporations would rather pay a couple of grand to replace a server than loose the data thereon (which is probably more). But really, nothing on my hard disk is absolutely crucial. The worst is the time it took me to get my linux installation fine tuned, but that's not too bad either. I can't afford having to replace a CPU or CPU and board. I don't think it will blow up in smoke if the fan stops - I've got a fairly substantial heatsink for a 1800+ and fans setup so there's a good solid flow of air through the case. But I'd rather be safe than sorry. I don't mind pulling the plug if I'm in front of the machine and I see the temperature rize sharply (I'm really keeping an eye on that), but I need something that would take care of this for me when I'm AWAY. For what it's worth, as far as data protections goes: I'm using ReiserFS. Where I live, power isn't exactly too stable, so short power outages aren't uncommon. I've had them happen while I'm working on my computer, and I've never lost anything because of that. Thanks again Hans
and when the machine is connected to an adequate UPS power is not an issue. Fire or other natural disaster shouldn't be an issue here - if you are running to save your life I hope you aren't stopping in the computer room on the way out!
Your opinion that your hardware is more valuable than your data is quite different than any other opinion I've heard. I suppose it begs the question - why store the data anyplace if it isn't that valuable? Hardware can always be had relatively cheaply, recreating data isn't free unless your time has no value (to coin an old phrase).
-- John LeMay KC2KTH Senior Enterprise Consultant NJMC | http://www.njmc.com | Phone 732-557-4848 Specializing in Microsoft and Unix based solutions
-- Check the headers for your unsubscription address For additional commands send e-mail to suse-linux-e-help@suse.com Also check the archives at http://lists.suse.com Please read the FAQs: suse-linux-e-faq@suse.com
== Download ringtones, logos and picture messages at Ananzi Mobile Fun. http://www.ananzi.co.za/cgi-bin/goto.pl?mobile
On Sunday 16 March 2003 20:43, H du Plooy wrote:
This still does the stopping_this_service stopping_that_service and takes a good 30 seconds before the machine switches off. This is too long.
I want it to get the "switch-off-now!" signal, unmount all filesystems regardless of weather they're being used or not, and cut the power. Or better - cut the power straight away. This is emergency procedure. My data is important but NOT as valuable as the hardware.
Don't you have an emergency shutdown feature in your motherboard BIOS? I think most motherboards with heat sensors also have that feature to emergency kill the system when it reaches a preset max temperature I don't know of a way to automate it from within linux though, but it should be possible, since you can poweroff the system instantly with the magic sysrq key 'o'. I haven't seen it implemented in a program though.
On Sun, 16 Mar 2003 22:08:40 +0100
Anders Johansson
Don't you have an emergency shutdown feature in your motherboard BIOS? I think most motherboards with heat sensors also have that feature to emergency kill the system when it reaches a preset max temperature
Oddly not. It's a Gigabyte GA-7ZXE. All the Gigabyte boards with seonsors on I've ever worked with had the feature. This one just reports. It might be because the BIOS I have (FAe) on is considered "alpha." It's the only one that sees my CPU correctly though. With any other bios version on, the machine just doesn't power up. I'm keeping track of bios updates and loads them as they become availale. I don't see any difference from one to the next, but I'm hoping for more features. I've written to Gigabyte about this too - I hope they implement it! Thanks Hans == Download ringtones, logos and picture messages at Ananzi Mobile Fun. http://www.ananzi.co.za/cgi-bin/goto.pl?mobile
participants (6)
-
Anders Johansson
-
Curtis Rey
-
Gary Traffanstedt
-
H du Plooy
-
John LeMay
-
Tom Emerson