Hi All,
I know Centos 5 is almost EOL and all this is old but....
I'm helping a colleague who has moved a Centos 5 install on an IDE disk to system with a Foxconn M'board. His idea is to use the SATA interface that's on the Foxconn.
The thing boots OK and then gets into a loop trying to start the interfaces to the SATA disks. Oceans of stuff like this in /var/log/messages
Aug 7 04:23:32 filestore kernel: ata1: hard resetting link Aug 7 04:23:32 filestore kernel: ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 310) Aug 7 04:23:32 filestore kernel: ata1.00: configured for UDMA/33 Aug 7 04:23:32 filestore kernel: ata1.01: configured for UDMA/33 Aug 7 04:23:32 filestore kernel: ata1: EH complete Aug 7 04:24:02 filestore kernel: ata1.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen Aug 7 04:24:02 filestore kernel: ata1.00: cmd 25/00:08:00:88:e0/00:00:e8:00:00/e0 tag 0 dma 4096 in Aug 7 04:24:02 filestore kernel: res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout) Aug 7 04:24:02 filestore kernel: ata1.00: status: { DRDY }
I have limited the SATA speed to 1.5G with a kernel param in Grub to see if that helps.
After a few hours the thing seems to settle down and can see this
cat /proc/partitions | grep sd
8 0 1953514584 sda 8 16 1953514584 sdb 8 32 1953514584 sdc
But the devices are non-functional. eg:-
smartctl -a /dev/sda
Smartctl open device: /dev/sda failed: No such device
and
fdisk /dev/sda
Unable to open /dev/sda
So I reckon the JMicron JMB361 controller [02:00.0 SATA controller: JMicron Technology Corp. JMB361 AHCI/IDE (rev 02)] is not something that Centos 5 is very keen on.
Its loaded the ahci module for it
uname says
Linux filestore 2.6.18-411.el5 #1 SMP Mon Jul 11 18:16:41 CDT 2016 i686 i686 i386 GNU/Linux
This thread -> https://bugzilla.redhat.com/show_bug.cgi?id=462425 <- suggests that a fix may have been backported from 2.6.27
Does this look familiar to anyone? And is there a remedy or would I be best to swap in a better SATA controller in there? Suggestions??
Thanks
Ken