Hello Gurus,
I have recently migrated a Solaris 8 (E3500) server to Solaris 10 (Sun Fire
V440). On this server we have an Overland REO 9000 iSCSI unit attached with
gigabit nic's.
My problem today is that my performance is extremly bad when switched from Sol
8. It has taken me the better half of the weekend to create a file system of
50Gb just using newfs on the device.
Example, it took me ~46minutes to create a 2Gb 'empty' file using a small
C-script.
[EMAIL PROTECTED] tmp]# time ./prog_disk -c 131072 -o
/export/home/reodata/michelin_2G_21 16384
Pid : 3040
131072 blks de 16384 car (2789.0 s): 2048.00 Mo 752 Ko/s - avant fsync
131072 blks de 16384 car (2799.4 s): 2048.00 Mo 749 Ko/s - apres fsync
real 46m43.133s
user 0m0.538s
sys 0m37.699s
On our HP machines same creation takes ~2 min on the same iSCSI box with
gigabit nic's.
HP/UX
epshpb:/ # time ./prog_disk -c 131072 -o
/export/home/doctracker3/michelin_2G_21 16384
Pid : 5210
131072 blks de 16384 car (111.6 s): 2048.00 Mo 18784 Ko/s - avant fsync
131072 blks de 16384 car (139.5 s): 2048.00 Mo 15031 Ko/s - apres fsync
real 2:19.5
user 0.1
sys 16.5
This is the output from current newfs running to create a 25Gb slice instead of
50Gb.
[EMAIL PROTECTED] /]# time newfs /dev/rdsk/c6t1d0s0
newfs: /dev/rdsk/c6t1d0s0 last mounted as /reo-data
newfs: construct a new file system /dev/rdsk/c6t1d0s0: (y/n)? y
Warning: 4096 sector(s) in last cylinder unallocated
/dev/rdsk/c6t1d0s0: 52428800 sectors in 8534 cylinders of 48 tracks, 128
sectors
25600.0MB in 534 cyl groups (16 c/g, 48.00MB/g, 5824 i/g)
super-block backups (for fsck -F ufs -o b=#) at:
32, 98464, 196896, 295328, 393760, 492192, 590624, 689056, 787488, 885920,
Initializing cylinder groups:
.. <--here i have 2 dot's so it is somehow working
/var/adm/messages
Feb 19 12:56:51 epsu82 iscsi: [ID 749434 kern.info] NOTICE: iscsi persistent
store update ok now filename:/etc/iscsi/iscsi_v1.dbc
Feb 19 13:00:01 epsu82 scsi: [ID 107833 kern.warning] WARNING: /iscsi/[EMAIL
PROTECTED],0 (sd1):
Feb 19 13:00:01 epsu82 SCSI transport failed: reason 'tran_err': retrying
command
[...]
Feb 19 13:25:34 epsu82 scsi: [ID 107833 kern.warning] WARNING: /iscsi/[EMAIL
PROTECTED],0 (sd1):
Feb 19 13:25:34 epsu82 SCSI transport failed: reason 'tran_err': giving up
Feb 19 13:25:34 epsu82 scsi: [ID 107833 kern.warning] WARNING: /iscsi/[EMAIL
PROTECTED],0 (sd1):
Feb 19 13:25:34 epsu82 SCSI transport failed: reason 'tran_err': giving up
Still running as you can see.
I have been in contact with SUN Support without being able to solve this
problem. I have attached latest patch cluster, kernel patch and iscsi patch. I
am very helpful for any light that can be lit on this problem.
Regards,
Pierre
This message posted from opensolaris.org
_______________________________________________
storage-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/storage-discuss
