Hi,
I currently have some trouble with LUNs exposed by Synology boxes.
As of Linux kernel 3.19, the max_sectors_kb value always is 32767 for
any LUN, which results into errors on the Synology target side and
reconnects on the Linux initiator side:
On the Synology target:
iSCSI: Unable to allocate memory for iscsi_cmd_t->iov_data.
iSCSI: iSCSI: Close -
I[iqn.1993-08.org.debian:01:99bdf6552818][192.168.xxx.78],
T[iqn.2000-01.com.synology:yyy.mylun.7e92dd8012]
iSCSI: iSCSI: Single CHAP security negotiation completed sucessfully.
iSCSI: iSCSI: Login -
I[iqn.1993-08.org.debian:01:99bdf6552818][192.168.xxx.78],
T[iqn.2000-01.com.synology:yyy.mylun.7e92dd8012][192.168.xxx.91:3260]
On the Linux initiator side:
Kernel reported iSCSI connection 4:0 error (1020 -
ISCSI_ERR_TCP_CONN_CLOSE: TCP connection closed) state (3)
connection4:0: detected conn error (1020)
iscsid: connection4:0 is operational after recovery (1 attempts)
In some cases, e.g. when stacking a LUN with LUKS encryption, I also see
a lot of I/O errors at the client side (in this example, I created an
ext3 file-system on it):
sd 11:0:0:1: [sdj] UNKNOWN(0x2003) Result: hostbyte=0x0e driverbyte=0x00
sd 11:0:0:1: [sdj] CDB: opcode=0x2a 2a 00 00 00 1a 00 00 1e 08 00
blk_update_request: I/O error, dev sdj, sector 6656
Buffer I/O error on dev dm-0, logical block 64, lost async page write
Buffer I/O error on dev dm-0, logical block 65, lost async page write
Buffer I/O error on dev dm-0, logical block 66, lost async page write
Buffer I/O error on dev dm-0, logical block 67, lost async page write
Buffer I/O error on dev dm-0, logical block 68, lost async page write
Buffer I/O error on dev dm-0, logical block 69, lost async page write
Buffer I/O error on dev dm-0, logical block 70, lost async page write
Buffer I/O error on dev dm-0, logical block 71, lost async page write
Buffer I/O error on dev dm-0, logical block 72, lost async page write
Buffer I/O error on dev dm-0, logical block 73, lost async page write
sd 11:0:0:1: [sdj] UNKNOWN(0x2003) Result: hostbyte=0x0e driverbyte=0x00
sd 11:0:0:1: [sdj] CDB: opcode=0x2a 2a 00 3e 78 c5 50 00 40 00 00
blk_update_request: I/O error, dev sdj, sector 1048102224
...
buffer_io_error: 6835 callbacks suppressed
I suspect this is due to the fact that the Synology target does not
present any meaningful block limits, e.g.:
# sg_vpd -p bl /dev/sdj
Block limits VPD page (SBC):
Write same no zero (WSNZ): 0
Maximum compare and write length: 0 blocks
Optimal transfer length granularity: 0 blocks
Maximum transfer length: 0 blocks
Optimal transfer length: 0 blocks
Maximum prefetch length: 0 blocks
Maximum unmap LBA count: 0
Maximum unmap block descriptor count: 0
Optimal unmap granularity: 0
Unmap granularity alignment valid: 0
Unmap granularity alignment: 0
Maximum write same length: 0x0 blocks
What bothers me most is that "Maximum transfer length" is set to zero,
which from my understanding of the spec means that no limits apply,
hence the Linux initiator is correctly not applying any limit here.
However, at the target side, I can see the following limits:
For block-level LUNs:
# cat /sys/kernel/config/target/core/.../attrib/hw_max_sectors
128
# cat /sys/kernel/config/target/core/.../attrib/max_sectors
128
...and for file-level LUNs:
# cat /sys/kernel/config/target/core/.../attrib/hw_max_sectors
1024
# cat /sys/kernel/config/target/core/.../attrib/max_sectors
1024
According to a sector size of 512 bytes, I should be safe with 64
respectively 512 for the value of max_sectors_kb. I have tested
extensively with heavy load and different scenarios (e.g. formatting,
iozone single and multi-threaded tests) that after lowering that value
on the initiator side seems to resolve the problem.
However, while testing I also came up with the following limits for
max_sectors_kb:
fileio: 4096
iblock: 512
...which to me does not make much sense, except if the sector size would
be 4096, which it isn't (at least not reported):
$ cat /sys/block/sdj/queue/hw_sector_size
512
So, I am puzzled. Can someone help me making sense out of this?
Btw., the DI VPD of the LUNs are:
# sg_vpd -p di_lu /dev/sdj
Device Identification VPD page:
Addressed logical unit:
designator type: NAA, code set: Binary
0x600140520cc460dd155ed32b5db631d3
designator type: T10 vendor identification, code set: ASCII
vendor id: SYNOLOGY
vendor specific: iSCSI Storage:20cc460d-155e-32b5-b631-3e5d006efbd2
designator type: Logical unit group, code set: Binary
Logical unit group: 0x0
Is there a way to _configure_ attributes for specific SCSI devices?
Thanx for any help! :)
Regards,
Ancoron
--
You received this message because you are subscribed to the Google Groups
"open-iscsi" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/open-iscsi.
For more options, visit https://groups.google.com/d/optout.
