RE: DMA TIMEOUT
> Add the following to your kernel config: > makeoptions DEBUG=-g > options DDB, KDB, GDB > options INVARIANTS > options INVARIANT_SUPPORT > options WITNESS_KDB > options WITNESS_SKIPSPIN > # Add this if you're using a firewire console > options BREAK_TO_DEBUGGER,ALT_BREAK_TO_DEBUGGER > # Do an unattended dump and reboot > options KDB_UNATTENDED > > Then recompile your kernel. The problem you're going to run into > since the problems you're having seem to be with the ATA code is that > there is a good chance you're not going to be able to dump the crash > dump to disk. I'd highly recommend using a serial or firewire > console. The problem persisted so I rebuilt with the debugging options you gave me. Other than the firewire console option if a panic occurs where do I go to obtain a backtrace? Yes ado1a is dying with a complete failure it seems so I am guessing I am going to need to look into the firewire option. But I would like to look into the conventional backtrace retrieval first. -- Wil Hatfield ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
RE: DMA TIMEOUT
> > >> Did you use a 80- or 40-ATA cable? > >> If you've configured your drives to do UATA-66 or faster then > >> FreeBSD (or any other OS for that matter) will crash if > you connect a > >> second drive... This is new one to me! I guess my 3 x UATA-100 are super drives of some sort not to crash!! > > It's an 80 wire. I have two drives on nearly all of my machines and > > never had an issue with crashing until just recently. > Started in 5.4, > > gone in 6.1-PRE, back in 6.1-RC. It doesn't happen alot > though under > > 6.1-RC. But under 5.4 it was about every 8 hours on > average. So in one > > sense 6.1 is still saving my arsh. > > > > -- > > Wil Hatfield > > > > I suspect some kind of hardware problem, and not a software problem... > If you can, boot into another OS, preferbly windows, since it > will crash on just about anything, you can use your swap > partition to install it... I have had this case on one machine running FreeBSD 5.4, it would happen intermittently, but after replacing the power supply it appears to have resolved the issue. Also, once I noticed this problem with MegaRac G2 remote management card installed! However, when you've upgraded to FreeBSD 6.0, did you test the system extensively for long period of time before upgrading to 6.1-RC ? If it worked well in 6.0 and stopped working in 6.1, I've to lean towards SW issue assuming all things remained the same. Tamouh ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: DMA TIMEOUT
Anish Mistry wrote: On Thursday 13 April 2006 07:47, Wil Hatfield wrote: I had a similar situation under 6.0. My secondary drive would throw DMA read errors at bootup, adding several minutes to the boot process, so I ran it in PIO mode. The upgrade to 6.1 solved it, both drives work fine as DMA now. Looks like the DMA errors are back in 6.1-RC with ATA or at least similar DMA errors. The new one froze my machine with an error like "Error while performing DMA_WRITE command". A new twist to the WRITE_DMA Timeouts of 5.4. I am starting to think that they aren't going to get the ATA issues all worked out anytime soon so they are changing the errors. ;-) And of course no automatic reboot on panic. Do you have a backtrace? Will such a backtrace actually help resolve the no reboot problem? Or will it just provide information about the panic? --Alex ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: DMA TIMEOUT
On Thu, 13 Apr 2006 21:57:58 -, Wil Hatfield <[EMAIL PROTECTED]> wrote: Did you use a 80- or 40-ATA cable? If you've configured your drives to do UATA-66 or faster then FreeBSD (or any other OS for that matter) will crash if you connect a second drive... It's an 80 wire. I have two drives on nearly all of my machines and never had an issue with crashing until just recently. Started in 5.4, gone in 6.1-PRE, back in 6.1-RC. It doesn't happen alot though under 6.1-RC. But under 5.4 it was about every 8 hours on average. So in one sense 6.1 is still saving my arsh. -- Wil Hatfield I suspect some kind of hardware problem, and not a software problem... If you can, boot into another OS, preferbly windows, since it will crash on just about anything, you can use your swap partition to install it... ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: DMA TIMEOUT
On Thursday 13 April 2006 18:01, Wil Hatfield wrote: > > Do you have a backtrace? > > No. To be honest I have worked a little with the debugging but that > was a long time ago. We ran 4.10 for eons and never needed to debug > a thing. I got spoiled I guess. So now that I need to add debugging > and backtrace to my arsenal of knowledge could someone point me in > the direction of a howto? It would be good to catch up. Add the following to your kernel config: makeoptions DEBUG=-g options DDB, KDB, GDB options INVARIANTS options INVARIANT_SUPPORT options WITNESS_KDB options WITNESS_SKIPSPIN # Add this if you're using a firewire console options BREAK_TO_DEBUGGER,ALT_BREAK_TO_DEBUGGER # Do an unattended dump and reboot options KDB_UNATTENDED Then recompile your kernel. The problem you're going to run into since the problems you're having seem to be with the ATA code is that there is a good chance you're not going to be able to dump the crash dump to disk. I'd highly recommend using a serial or firewire console. To setup a serial console just add the following to your /boot/loader.conf or set it at the loader prompt. console="comconsole" -- Firewire -- For a firewire console you'll need to load the following modules assuming they aren't in your kernel. dcons_load="YES" dcons_crom_load="YES" You may need to do a fwcontrol -r on both systems to make sure everything is in sync. Once loaded on your client machine, just do a fwcontrol to get the address of the other system. Then to connect to the server just: dconschat -t 00-00-0e-10-00-b0-29-d0 Where 00-00-0e-10-00-b0-29-d0 is the firewire address of the server. -- Anish Mistry pgp6EGGkx3C7h.pgp Description: PGP signature
RE: DMA TIMEOUT
> Do you have a backtrace? No. To be honest I have worked a little with the debugging but that was a long time ago. We ran 4.10 for eons and never needed to debug a thing. I got spoiled I guess. So now that I need to add debugging and backtrace to my arsenal of knowledge could someone point me in the direction of a howto? It would be good to catch up. -- Wil Hatfield ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
RE: DMA TIMEOUT
> Did you use a 80- or 40-ATA cable? > If you've configured your drives to do UATA-66 or faster then > FreeBSD (or > any other OS for that matter) will crash if you connect a second drive... It's an 80 wire. I have two drives on nearly all of my machines and never had an issue with crashing until just recently. Started in 5.4, gone in 6.1-PRE, back in 6.1-RC. It doesn't happen alot though under 6.1-RC. But under 5.4 it was about every 8 hours on average. So in one sense 6.1 is still saving my arsh. -- Wil Hatfield ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: DMA TIMEOUT
On Thu, 13 Apr 2006 15:53:57 -, Anish Mistry <[EMAIL PROTECTED]> wrote: On Thursday 13 April 2006 07:47, Wil Hatfield wrote: > I had a similar situation under 6.0. My secondary drive would > throw DMA read > errors at bootup, adding several minutes to the boot process, so > I ran it in > PIO mode. The upgrade to 6.1 solved it, both drives work fine as > DMA now. Looks like the DMA errors are back in 6.1-RC with ATA or at least similar DMA errors. The new one froze my machine with an error like "Error while performing DMA_WRITE command". A new twist to the WRITE_DMA Timeouts of 5.4. I am starting to think that they aren't going to get the ATA issues all worked out anytime soon so they are changing the errors. ;-) And of course no automatic reboot on panic. Do you have a backtrace? Did you use a 80- or 40-ATA cable? If you've configured your drives to do UATA-66 or faster then FreeBSD (or any other OS for that matter) will crash if you connect a second drive... ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: DMA TIMEOUT
On Thursday 13 April 2006 07:47, Wil Hatfield wrote: > > I had a similar situation under 6.0. My secondary drive would > > throw DMA read > > errors at bootup, adding several minutes to the boot process, so > > I ran it in > > PIO mode. The upgrade to 6.1 solved it, both drives work fine as > > DMA now. > > Looks like the DMA errors are back in 6.1-RC with ATA or at least > similar DMA errors. The new one froze my machine with an error like > "Error while performing DMA_WRITE command". A new twist to the > WRITE_DMA Timeouts of 5.4. I am starting to think that they aren't > going to get the ATA issues all worked out anytime soon so they are > changing the errors. ;-) > > And of course no automatic reboot on panic. Do you have a backtrace? -- Anish Mistry pgp623nOlpwIP.pgp Description: PGP signature
RE: DMA TIMEOUT
> I had a similar situation under 6.0. My secondary drive would > throw DMA read > errors at bootup, adding several minutes to the boot process, so > I ran it in > PIO mode. The upgrade to 6.1 solved it, both drives work fine as DMA now. Looks like the DMA errors are back in 6.1-RC with ATA or at least similar DMA errors. The new one froze my machine with an error like "Error while performing DMA_WRITE command". A new twist to the WRITE_DMA Timeouts of 5.4. I am starting to think that they aren't going to get the ATA issues all worked out anytime soon so they are changing the errors. ;-) And of course no automatic reboot on panic. -- Wil Hatfield ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: DMA TIMEOUT
On Thursday 13 April 2006 01:58, Perttu Laine wrote: > I have problem with ATA drive having DMA TIMEOUTs. > It might work day or two fine with DMA mode on but then system halts on > those dma timeout problems. > > hardware is: > [EMAIL PROTECTED]:5:0: class=0x018085 card=0x4d68105a chip=0x4d69105a > rev=0x02 hdr=0x00 > vendor = 'Promise Technology Inc' > device = 'PDC20269 Ultra133 TX2 EIDE Controller' > class= mass storage > > and PDC20269 is on supported list (5.4-RELEASE). > > and has one 160GB drive in it. When I first got this problem I tried change > cable, changed drive to another similar one and moved drive to different > place on card. No help. Now it has been running on PIO mode about a year > without any problems at all. > > So I wonder if there actually is some problem in ata drivers and is there > maybe any fixes on those drivers in new releases? So would upgrade to 5.5 > or 6.1 help on this? upgrade from 5.3 to 5.4 didn't help... > > I can actually run it on PIO-mode but sometimes DMA could give little more > speed so if you would be cool if I can get it work with DMA mode on. I had a similar situation under 6.0. My secondary drive would throw DMA read errors at bootup, adding several minutes to the boot process, so I ran it in PIO mode. The upgrade to 6.1 solved it, both drives work fine as DMA now. David -- Sure God created the world in only six days, but He didn't have an established user-base. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"