Hi guys.
I am in the progress of copying data between two of my linux machines and its basically just not working. I am getting transfers of 2-3Mb/s over the network (using Samba).
Both machines are running Suse 9.3 and have two network cards each, one 100Mbit and one 1000Mbit. It does not matter which of the cards I use. The problem seems to be the source host, where the IO wait is usally >60 while the destination host is basically idle. And copying to the dest. host from WinXP I get transfers around 15-20MB/s when using the 1000Mbit card on the dest. host.
Copying locally between disks or to USB drive on the source host seems to be fine so it is not the disk IO that is a problem.
I know the 1000Mbit equipment I have is not exactly high end (RTL 8169) but 2-3MB/s over gigabit ethernet is a little frustrating.
Has anyone had similar issues and perhaps some tips on how to fix this ?
Maybe its time to dig up those Intel 1000/F cards I have around here somewhere and go fiber. Maybe they fixed the bug on all 1000Mbit Intel cards which reset the device every 2 seconds.
ip
/*Network copy : source machine*/
/*Network copy : destination machine*/
/*Disk 2 Disk copy : source machine*/
/*Disk 2 USB drive copy : source machine*/
/*WinXP to dest. host copy : destination machine */
I am in the progress of copying data between two of my linux machines and its basically just not working. I am getting transfers of 2-3Mb/s over the network (using Samba).
Both machines are running Suse 9.3 and have two network cards each, one 100Mbit and one 1000Mbit. It does not matter which of the cards I use. The problem seems to be the source host, where the IO wait is usally >60 while the destination host is basically idle. And copying to the dest. host from WinXP I get transfers around 15-20MB/s when using the 1000Mbit card on the dest. host.
Copying locally between disks or to USB drive on the source host seems to be fine so it is not the disk IO that is a problem.
I know the 1000Mbit equipment I have is not exactly high end (RTL 8169) but 2-3MB/s over gigabit ethernet is a little frustrating.
Has anyone had similar issues and perhaps some tips on how to fix this ?
Maybe its time to dig up those Intel 1000/F cards I have around here somewhere and go fiber. Maybe they fixed the bug on all 1000Mbit Intel cards which reset the device every 2 seconds.
ip
/*Network copy : source machine*/
Code:
linux1:~ # vmstat 1 1000
procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu----
r b swpd free buff cache si so bi bo in cs us sy id wa
0 6 2760 3484 71232 67764 0 0 6200 0 4776 2835 2 21 10 [B]67[/B]
1 1 2760 3112 72608 67404 0 0 9548 80 5802 3141 3 28 0 [B]69[/B]
0 5 2760 3012 72192 66976 0 0 7060 32 5878 3298 2 29 5 [B]64[/B]/*Network copy : destination machine*/
Code:
ntop:~ # vmstat 1 1000
procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu----
r b swpd free buff cache si so bi bo in cs us sy id wa
0 0 0 13636 407980 515392 0 0 43 1279 1906 1309 1 3 94 2
0 0 0 13876 407492 515772 0 0 0 8244 7410 9200 1 21 77 1
0 0 0 13316 407848 515948 0 0 0 8432 4741 5260 1 15 81 3
0 0 0 13316 407364 516656 0 0 0 8244 6047 7145 0 19 80 1/*Disk 2 Disk copy : source machine*/
Code:
linux1:/ # vmstat 1 1000
procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu----
r b swpd free buff cache si so bi bo in cs us sy id wa
0 3 2760 3224 67180 72616 0 0 42508 19820 1508 1342 5 70 0 25
0 5 2760 3072 65544 74432 0 0 41812 21784 1313 1173 1 74 0 25
1 1 2760 3016 65848 74308 0 0 48404 23660 1363 1217 0 82 0 18/*Disk 2 USB drive copy : source machine*/
Code:
linux1:/ # vmstat 1 1000
procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu----
r b swpd free buff cache si so bi bo in cs us sy id wa
4 3 2996 4048 69528 66248 0 24 43672 23040 2156 2781 3 74 0 23
2 3 2996 3280 69036 67536 0 24 44316 22172 2128 2825 4 73 0 23
0 5 2996 3404 68700 67248 0 0 43588 22516 2134 2825 2 76 0 22/*WinXP to dest. host copy : destination machine */
Code:
ntop:~ # vmstat 1 1000
procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu----
r b swpd free buff cache si so bi bo in cs us sy id wa
5 1 0 10304 433004 493108 0 0 0 93656 1711 1553 0 100 0 0
1 0 0 13380 423572 499584 0 0 0 352 23762 28458 1 28 69 2
1 1 0 13016 435036 485136 0 0 0 65824 10315 12515 0 68 17 15
Last edited:
