"SATA link down" on non-existing sata channel
I run SuSE 11.1 with kernel 2.6.27.10 and basically everything works fine. So why post here?
Well, I noticed that my var/log/messages started growing very rapidly and discovered that my kernel is pumping out SATA channel errors. Code:
Jan 4 22:11:00 lambda kernel: ata10: exception Emask 0x10 SAct 0x0 SErr 0x4000000 action 0xe frozen Looking in the boot.msg I can see that the physical 3 drives on the first three out of 6 SATA channels are in fact numbered from 3 and up and seemingly correctly setup as Code:
<6>ata3: SATA max UDMA/133 abar m2048@0xf7ffc000 port 0xf7ffc100 irq 216 Can anyone suggest a good next step to pinpoint this (and perhaps point to some ATA channel info so I can read up some more on how this actually works)? My lspci looks like Code:
00:00.0 Host bridge: Intel Corporation X58 I/O Hub to ESI Port (rev 12) Can anyone suggest a good next step to pinpoint this (and perhaps point to some ATA channel info so I can read up some more on how this actually works)? Rgds Johan |
Wow! I just bought a new S-ATA 160 gb hard drive + I'm downloading Open SuSE 11.1.
Was it a bad idea? Does SuSE 11.1. works fine with S-ATA or doesn't it? :( |
Possible caused by badly connected SATA wires
Since I remembered that this problem only started after I installed the third HDD I was wondering if I somehow had disturbed some cables by mistake.
I took the server apart again and made very sure all SATA cables, USB cables etc were firmly attached and rebooted. The problem now has disappeared. However, I'm still very curious to what device/unit/block devide etc. the sata10 channel was assigned to (since I don't have more than 6+1 real SATA channels). Is it possible for some USB block devices (like SM/MMC card reader) to show up masqueraded as a SATA channel? I just don't know enough about the kernel architecture regarding handling of block/HDD/SATA devices to really now in detail what really was going on. Regarding your question on SuSE 11.1 I would say that in general it is a fairly good release and the issues most people have seen is when they use KDE 4.1 (even though SuSE has backported a fair amount of fixes from the 4.2). Remember though that if you use it as a server (perhaps running a IMAP server that uses the inotify mechanism) you will want to think twice since the standard kernel shipped 2.6.27.7 has a serious bug in the inotify handling that can cause un-stoppable runaway processes that requires a hard reboot of the server. You need to upgrade to a 2.6.27.10 kernel yourself (or wait until SuSE has a suitable update) Johan |
Sorry for me being lazy for not searching this online but since you seem to know: which KDE comes up with SuSE 11.1?
|
too quick to declare victory
.. actually I was a bit too quick to declare that this problem is solved. It apparently had nothing to do with loose cables.
After ~4 minutes after the logs starts too fill up again every 2s with blocks of Code:
Jan 4 23:22:03 lambda kernel: ata10: exception Emask 0x10 SAct 0x0 SErr 0x4000000 action 0xe frozen Johan |
All times are GMT -5. The time now is 05:54 PM. |