Hello, from time to time my FreeBSD freezes under heavy hdd load and only a hard reset will bring it back to life with fsck complaining about 'Softupdate Inconsistencies'. I had this behaviour on 5.3-RELEASE, 5.3-RELEASE-p5 and now i have it on 5.4-PRERELEASE. I am using a custom kernel with SMP enabled on a P4 3.2GHz for hyperthreading. One hdd is a SATA drive and it acts fine. The other hdd however is an ATA133 drive and i suspect it to be the problem, since freezes only occur when it is busy (eg. copying much data from a DVD/HDD to it or compiling a port). Whenever the system freezes there is no warning or log entry at all. I used 'smartmontools' to check the drive, but there was not found anything and the hdd appeared to be fully operational. Could the freezes come from a faulty IDE hdd (which would mean that I better get rid of it), or are there other possiblities. Thank you, Andreas Rudisch -- GnuPG key : 0xD25FCC81 | http://cyb.websimplex.de/pubkey.asc Fingerprint: D182 6F22 7EEC DD4C 0F6E 564C 691B 0372 D25F CC81 -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 187 bytes Desc: This is a digitally signed message part Url : http://lists.freebsd.org/pipermail/freebsd-stable/attachments/20050307/fb8ae675/attachment.bin
On Monday 07 March 2005 09:38 am, cyb wrote:> from time to time my FreeBSD freezes under heavy hdd load and only a > hard reset will bring it back to life with fsck complaining about > 'Softupdate Inconsistencies'.I had similar issues on an Athlon machine under 5.3. In my case it turned out that the CPU was running extremely hot. It would frequently freeze while portupgrade was backing up old versions of a port. I initially suspected a disk problem as well, but I haven't yet had any problems since cleaning my CPU and chipset fans. On reflection, my freezes were probably due to the CPU running hotter than usual when running bzip2 (which I believe is how portupgrade stores its backups, and is very CPU-intensive). It is possible that your drive is the culprit, but you would probably get console messages about I/O failures rather than just a hard freeze. Flaky memory or power supply are also possibilities. JN
Hi. On Mon, Mar 07, 2005 at 05:38:29PM +0100, cyb wrote:> Could the freezes come from a faulty IDE hdd (which would mean that I > better get rid of it), or are there other possiblities.How much RAM do you have? I also had a freezing problem on a machine with 4 GB of RAM. As soon as I reduced the RAM to 2 GB the machine became stable. I've played around with KVA_PAGES, vm.kmem_size_max and could only influence the time it took to crash the machine. - Oliver -- | Oliver Brandmueller | Offenbacher Str. 1 | Germany D-14197 Berlin | | Fon +49-172-3130856 | Fax +49-172-3145027 | WWW: http://the.addict.de/ | | Ich bin das Internet. Sowahr ich Gott helfe. | | Eine gewerbliche Nutzung aller enthaltenen Adressen ist nicht gestattet! | -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 187 bytes Desc: not available Url : http://lists.freebsd.org/pipermail/freebsd-stable/attachments/20050307/b7d395ea/attachment.bin
On Mon, 2005-03-07 at 09:54 -0700, John Nielsen wrote:> Flaky memory or power supply are also possibilities.On Mon, 2005-03-07 at 18:10 +0100, Oliver Brandmueller wrote:> How much RAM do you have? I also had a freezing problem on a machine > with 4 GB of RAM. As soon as I reduced the RAM to 2 GB the machine > became stable. I've played around with KVA_PAGES, vm.kmem_size_max > and > could only influence the time it took to crash the machine.I have 1GB (2x512MB) PC3200 DDR memory, memtest86 hasn't found anything. Andreas -- GnuPG key : 0xD25FCC81 | http://cyb.websimplex.de/pubkey.asc Fingerprint: D182 6F22 7EEC DD4C 0F6E 564C 691B 0372 D25F CC81 -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 187 bytes Desc: This is a digitally signed message part Url : http://lists.freebsd.org/pipermail/freebsd-stable/attachments/20050307/113d1101/attachment.bin
On Thu, 10 Mar 2005 11:04:50 -0800, Doug Hardie <bc979@lafn.org> wrote:> You might want to install the smartmontools port and use it to access > the disk performance data. That should give you good information on > the status of the drives and any problems they have encountered. >Smartmontools reports no problems with the drives; the system doesn't appear to be over heating at all. A bit more investigation shows it's panic'ing with "privileged instruction in non-priviliged mode" in what appears to be lseek(). This may or may not be related to the DMA TIMEOUT problem as I didn't see any timeouts before any of the panics I've witnessed. A bit more reading around suggests that the timeout issue is strongly related to my SI3112 SATA controller so I'm thinking of changing it; can anyone please recommend an SATA controller they're using which hasn't given them any problems? (Even better if they have Maxtor drives hanging off it!) I am also attempting to upgrade my system to a newer cvs but unfortunately I can't buildkernel on it - it'll either panic or generate enough timeouts to make "make" fail. If anyone could point me at a HOWTO or similar on cross-compiling then I'll compile it on my other machine instead (2xP3-1Ghz vs my desktop being an Athlon XP 1700+). All help appreciated as currently my desktop is becoming unusable and I fear for the safety of my data with the sheer number of panics being generated by disk use :)