On Fri, 09 Apr 2004, martin f krafft wrote: > also sprach martin f krafft <[EMAIL PROTECTED]> [2004.04.09.0034 +0200]: > > > No. nmi_watchdog=1 means IOAPIC, nmi_watchdog=2 means LAPIC. You > > > need to have the NMI Watchdog compiled into the kernel. > > > > This is my problem. I have looked around the LAPIC stuff and in the > > Watchdog section, but there is nothing about an NMI watchdog. I have > > enabled the Hangcheck timer and the Software watchdog. > > Okay, I did find it -- kernel/Documentation/nmi_watchdog.txt told me > all that I needed. > > Last question here: is it adviseable to run the NMI watchdog on > a productive system? After all, uptime is important, and as far as
It won't hurt. But nmi_watchdog is only for usermode and kernemode hangs. The NMI watchdog is useless against nasty bugs (hw or sw) that make the hardware unstable. I run my machine with it enabled, since if it ever decides to hang, I want an OOPs to have a slim chance of being produced before the TCO watchdog kicks in. I get about 101 NMIs per second on each CPU using nmi_watchdog=1. HZ=100 in this machine, I suppose... This is a not-really-SMP machine, with a single P4 HT processor. The software watchdog will reboot your machine (and AFAIK it might very well be using NMIs to do it, too). If you can use a chipset watchdog, however, that's much better (e.g. the TCO timers in most Intel systems, and _especially_ the IPMI watchdog in servers with a baseboard controller worth something). You can only have one watchdog listening on /dev/watchdog, btw. > I understand, the NMI watchdog should be able to reboot the machine > more often than it happens without the NMI watchdog. Of course, > I won't stop until this machine is rock stable. But in the mean > time... ? It might help, yes. But only if the hang is not killing the CPU or the northbridge. Otherwise, the CPU cannot execute the NMI code :) and any software watchdog will be useless. -- "One disk to rule them all, One disk to find them. One disk to bring them all and in the darkness grind them. In the Land of Redmond where the shadows lie." -- The Silicon Valley Tarot Henrique Holschuh -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
