Red Hat Bugzilla – Bug 430560
system hangs after a few hours of uptime (nForce3 ADMA 4GB problem?)
Last modified: 2008-06-27 05:47:42 EDT
Description of problem:
After a cold boot the system will hang completely after a few hours. This seems
to happen faster when there is disk activity. Investigating the problem suggest
a problem with the nForce3 SATA controller in ADMA mode and 4GB of memory (?)
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Cold boot system
2. Mildly exercise the system disk (ie. download some files)
3. System hangs, reset or hard boot needed
System hangs solid, reset or hard boot needed to recover
System running for weeks on end without problems
This system ran fine for weeks on end on FC6 x86_64 using the latest FC6 kernel.
Disabling the SATA ADMA function does not seem to work (/etc/modprobe.conf ->
options sata_nv adma=0) or using boot options sata_nv.adma=0 (Unknown boot
option `sata_nv.adma=0': ignoring.
This system has 4GB of memory.
Created attachment 293186 [details]
Relevant ata bits from dmesg
Relevant bits from dmesg. smartctl short and long test on the disk gives no
(In reply to comment #0)
> Disabling the SATA ADMA function does not seem to work (/etc/modprobe.conf ->
> options sata_nv adma=0) or using boot options sata_nv.adma=0 (Unknown boot
> option `sata_nv.adma=0': ignoring.
You need to rebuild the initrd after disabling adma in modprobe.conf. (Or just
uninstall and reinstall the kernel.)
OK, after finding my server crashed again (disk led permanently on) I rebuild
the initrd and rebooted. ADMA seems to be disabled now, let's see if it helps.
BTW is this a know bug?
Disabling ADMA definitely helps. The server has been running solid for 3 days now.
As this bug is not being picked up, it can be closed. With ADMA disabled and
running kernel 126.96.36.199-92.fc8 x86_64 (after a reinstall of F8):
uptime is: 11:44:47 up 39 days, 23:02