ata reset when reading from Raid 1 in the Guest

[jakap]

hi

i want to use a local ecrypted Raid 1 for storing data. So i used "ide1: /dev/md1" in the config file. The harddisk appears in the Guest and i decrypt it using cryptsetup.
Read/Write speed is okay.

For testing propose i used dd to copy ~ 10GB from /dev/zero, that was okay. But then copying from the 10GB file to /dev/null for reading tests, the guest hangs up, in the host this appears in the syslog

ata5.00: exception Emask 0x0 SAct 0x1 SErr 0x0 action 0x0
Mar 22 17:33:17 server kernel: ata5.00: irq_stat 0x40000008
Mar 22 17:33:17 server kernel: ata5.00: failed command: READ FPDMA QUEUED
Mar 22 17:33:17 server kernel: ata5.00: cmd 60/00:00:20:2c:fc/02:00:3f:00:00/40 tag 0 ncq 262144 in
Mar 22 17:33:17 server kernel:         res 41/40:7e:a2:2c:fc/06:01:3f:00:00/40 Emask 0x409 (media error) <F>
Mar 22 17:33:17 server kernel: ata5.00: status: { DRDY ERR }
Mar 22 17:33:17 server kernel: ata5.00: error: { UNC }
Mar 22 17:33:17 server kernel: ata5.00: configured for UDMA/133
Mar 22 17:33:17 server kernel: ata5: EH complete

This occurs about 10 times. I think the Harddisk is okay, i made the same test in the host and nothing happened. Now im checking it with e2fsck + badblocks.

At the moment i use only one 1.5TB disk for the Raid1, the second is set to "missing", because i wanted to test first.

Maybe it isn't a good idea to use a Raid for "passthrough"? Maybe better using the logical Volume?

Thanks

 

[jakap]

Im still having this issue.

Writing to the disk is no problem, but reading always ends up in this

(host dmesg)

ata5.00: failed command: READ FPDMA QUEUED
ata5.00: cmd 60/08:00:78:de:18/00:00:40:00:00/40 tag 0 ncq 4096 in
         res 41/40:03:7d:de:18/06:00:40:00:00/40 Emask 0x409 (media error) <F>
ata5.00: status: { DRDY ERR }
ata5.00: error: { UNC }
ata5.00: configured for UDMA/133
sd 4:0:0:0: [sdb] Unhandled sense code
sd 4:0:0:0: [sdb] Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE
sd 4:0:0:0: [sdb] Sense Key : Medium Error [current] [descriptor]
Descriptor sense data with sense descriptors (in hex):
        72 03 11 04 00 00 00 0c 00 0a 80 00 00 00 00 00
        40 18 de 7d
sd 4:0:0:0: [sdb] Add. Sense: Unrecovered read error - auto reallocate failed
sd 4:0:0:0: [sdb] CDB: Read(10): 28 00 40 18 de 78 00 00 08 00
end_request: I/O error, dev sdb, sector 1075371645
Buffer I/O error on device md1, logical block 134421455
ata5: EH complete

edit:
i've tested it now with external usb drive and the local drive; all works. Now i don't know if it's because of the raid or LVM (but my guest is using a lvm for root), could someone test a raid, passtrough the md device and mount it locally, then copy ~2 GB from /dev/zero to the disk, then copy it into /dev/null.

What happens?

Best regards

 

[jakap]

Again me:

decrypting the Raid1 in the Host and using the lv for "ide1: /dev/mapper/lvShare" does also not work. Again ata Reset and then the Disk's DMA mode is off! Could someone test it please? Or do you have an idea, where this comes from?

Are there other drivers then the qemu harddisk driver?

Best regards..

 

[oeginc]

I'm not sure if my problem is related or not, but I too have been getting bus reset errors on 2 brand new server builds using the Highpoint 2640x4 RAID controller in a RAID-10 setup with 4 brand new seagate hard drives...

 

[jakap]

My problem was a defect harddisk i guess. It went back to the shop after several Reading Erros in a long Smart Test.

Can you test if there are problems without Proxmox? For example with a live cd?