Re: [pfSense Support] ACPI/APIC in loader.conf - watchdog timeouts
On Mon, Mar 23, 2009 at 1:02 AM, Dimitri Rodis dimit...@integritasystems.com wrote: Do you think this has any potential relevance to the firebox watchdog timeouts? Obviously I am going to test it and simply observe the results-- not too hard to reproduce the issue. It could. Also, there was a suggestion that using an SMP kernel would alleviate the issue also. Given that this is a single core P3, I don't know what difference it will make (obviously the kernel locking mechanisms are different), but is there a way to easily swap the kernel on embedded with an SMP version (if it isn't already--I don't know what the default is for an embedded image since there isn't an installer)? Mount it rw (run /etc/rc.conf_mount_rw) and copy over the kernel from a full install. Then switch back to ro with /etc/rc.conf_mount_ro and reboot. - To unsubscribe, e-mail: support-unsubscr...@pfsense.com For additional commands, e-mail: support-h...@pfsense.com Commercial support available - https://portal.pfsense.org
RE: [pfSense Support] ACPI/APIC in loader.conf - watchdog timeouts
So, the hint.apic.0.disabled=1 seems to have _significantly_ reduced the watchdog timeouts, but they are not completely gone, and the ones that are happening now seem to happen somewhat randomly. Browsing through the GUI does not seem to cause issues any more. I will continue with the SMP kernel testing tomorrow. Dimitri Rodis Integrita Systems LLC http://www.integritasystems.com -Original Message- From: cbuech...@gmail.com [mailto:cbuech...@gmail.com] On Behalf Of Chris Buechler Sent: Monday, March 23, 2009 6:05 PM To: support@pfsense.com Subject: Re: [pfSense Support] ACPI/APIC in loader.conf - watchdog timeouts On Mon, Mar 23, 2009 at 1:02 AM, Dimitri Rodis dimit...@integritasystems.com wrote: Do you think this has any potential relevance to the firebox watchdog timeouts? Obviously I am going to test it and simply observe the results-- not too hard to reproduce the issue. It could. Also, there was a suggestion that using an SMP kernel would alleviate the issue also. Given that this is a single core P3, I don't know what difference it will make (obviously the kernel locking mechanisms are different), but is there a way to easily swap the kernel on embedded with an SMP version (if it isn't already--I don't know what the default is for an embedded image since there isn't an installer)? Mount it rw (run /etc/rc.conf_mount_rw) and copy over the kernel from a full install. Then switch back to ro with /etc/rc.conf_mount_ro and reboot. - To unsubscribe, e-mail: support-unsubscr...@pfsense.com For additional commands, e-mail: support-h...@pfsense.com Commercial support available - https://portal.pfsense.org smime.p7s Description: S/MIME cryptographic signature
[pfSense Support] ACPI/APIC in loader.conf - watchdog timeouts
So I just came across this little tidbit while searching for potential solutions to the re: watchdog timeout issue on the firebox installs that I have pfSense running on. Some folks suggest that the problem is due to an interrupt storm which can result in a partial/total system hang. While doing further research, I found this: http://www.freebsd.org/doc/en/books/handbook/acpi-debug.html Specifically: -- 11.16.3.3 System Hangs (temporary or permanent) Most system hangs are a result of lost interrupts or an interrupt storm. Chipsets have a lot of problems based on how the BIOS configures interrupts before boot, correctness of the APIC (MADT) table, and routing of the System Control Interrupt (SCI). Interrupt storms can be distinguished from lost interrupts by checking the output of vmstat -i and looking at the line that has acpi0. If the counter is increasing at more than a couple per second, you have an interrupt storm. If the system appears hung, try breaking to DDB (CTRL+ALT+ESC on console) and type show interrupts. Your best hope when dealing with interrupt problems is to try disabling APIC support with hint.apic.0.disabled=1 in loader.conf. -- hint.apic.0.disabled=1? I thought it was hint.acpi.0.disabled=1 (see http://doc.pfsense.org/index.php/Booting_Options, and also the forum posts regarding firebox installs) Is there a typo here or are these two totally different things? I have not tried the hint.apic.0.disabled=1 yet, but I plan to tomorrow. Also, are the double quotes of particular importance? Some docs show them there, others don't. Any info appreciated.. I think these old end of life firebox x series units would be great for pfSense, provided we can get the watchdog timeouts to go away (and a specially sized sticker than can cover up the Firebox X logo J) Dimitri Rodis Integrita Systems LLC smime.p7s Description: S/MIME cryptographic signature
Re: [pfSense Support] ACPI/APIC in loader.conf - watchdog timeouts
On Mon, Mar 23, 2009 at 12:38 AM, Dimitri Rodis dimit...@integritasystems.com wrote: hint.apic.0.disabled=1? I thought it was hint.acpi.0.disabled=1 (see http://doc.pfsense.org/index.php/Booting_Options, and also the forum posts regarding firebox installs) APIC and ACPI are entirely different things. APIC is another one that can cause problems on some systems. http://en.wikipedia.org/wiki/Advanced_Programmable_Interrupt_Controller http://en.wikipedia.org/wiki/Advanced_Configuration_and_Power_Interface - To unsubscribe, e-mail: support-unsubscr...@pfsense.com For additional commands, e-mail: support-h...@pfsense.com Commercial support available - https://portal.pfsense.org
RE: [pfSense Support] ACPI/APIC in loader.conf - watchdog timeouts
Do you think this has any potential relevance to the firebox watchdog timeouts? Obviously I am going to test it and simply observe the results-- not too hard to reproduce the issue. Also, there was a suggestion that using an SMP kernel would alleviate the issue also. Given that this is a single core P3, I don't know what difference it will make (obviously the kernel locking mechanisms are different), but is there a way to easily swap the kernel on embedded with an SMP version (if it isn't already--I don't know what the default is for an embedded image since there isn't an installer)? Doing a full install on these fireboxes is pretty tough and requires some soldering (I believe) to get a keyboard header working, not to mention that you have to get the board completely out of the chassis to fit a video card on it. Thanks Chris.. Dimitri Rodis Integrita Systems LLC -Original Message- From: cbuech...@gmail.com [mailto:cbuech...@gmail.com] On Behalf Of Chris Buechler Sent: Sunday, March 22, 2009 9:44 PM To: support@pfsense.com Subject: Re: [pfSense Support] ACPI/APIC in loader.conf - watchdog timeouts On Mon, Mar 23, 2009 at 12:38 AM, Dimitri Rodis dimit...@integritasystems.com wrote: hint.apic.0.disabled=1? I thought it was hint.acpi.0.disabled=1 (see http://doc.pfsense.org/index.php/Booting_Options, and also the forum posts regarding firebox installs) APIC and ACPI are entirely different things. APIC is another one that can cause problems on some systems. http://en.wikipedia.org/wiki/Advanced_Programmable_Interrupt_Controller http://en.wikipedia.org/wiki/Advanced_Configuration_and_Power_Interface - To unsubscribe, e-mail: support-unsubscr...@pfsense.com For additional commands, e-mail: support-h...@pfsense.com Commercial support available - https://portal.pfsense.org smime.p7s Description: S/MIME cryptographic signature