Still unhappy with my network speed and I''ve tried all the usual tricks from this mailing list. When I to a virt-top it shows the network from and external machine to domU running at 64M/s, however the rsync is reporting the transfer working at 18M/s. After a few seconds virt-top drops to 25M/s and rsync to 8M/s. I''m assuming the difference is virt-top is reporting traffic coming into the device where rsync is reporting from end to end (through the virtual bridge) Would this be correct? Monitoring vmstat on the dom0 and domU *DomU:* procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------ r b swpd free buff cache si so bi bo in cs us sy id wa st 1 0 37388 191796 1372 19052 1 2 37 9 13 14 0 2 95 1 2 0 0 37388 191136 1380 19684 0 0 213 36 331 539 0 7 92 0 1 0 0 37388 190836 1388 19948 0 0 85 5 342 540 0 8 91 0 1 0 0 37388 190716 1388 20068 0 0 43 0 330 551 0 6 94 0 0 -> Started Transfer (About here) 1 0 37388 150576 1524 54632 0 0 279 6859 2719 2857 16 23 53 6 3 1 0 37388 77856 1600 123428 0 0 151 23521 4999 5647 42 50 1 0 7 1 0 37388 9036 1676 191660 0 0 43 23059 5011 5778 41 51 0 0 8 1 0 37388 3992 1300 195872 0 0 299 22208 4717 5630 39 52 0 0 9 1 0 37388 3708 1372 196904 0 0 272 23557 5122 5745 43 52 0 0 4 0 0 37388 197476 1376 13052 0 0 91 1381 721 973 2 10 88 0 0 0 0 37388 196948 1388 13564 0 0 172 29 328 549 0 7 93 0 0 -> Cancelled Transfer (About here) 0 0 37388 196948 1388 13704 0 0 43 0 317 540 0 7 93 0 0 *Dom0:* procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------ r b swpd free buff cache si so bi bo in cs us sy id wa st 0 0 102780 64920 30040 464004 0 1 2965 1137 10 3 1 1 92 1 5 0 0 102780 64648 30048 464016 0 0 47 889 744 1107 1 1 97 0 2 0 0 102780 64648 30068 464256 0 0 157 1084 979 1058 0 0 97 0 2 0 0 102780 64276 30072 464288 0 0 157 799 993 1076 0 0 96 0 4 1 0 102780 64276 30084 464288 0 0 88 319 566 948 0 1 95 0 4 -> Started Transfer (About here) 0 0 102780 64276 30088 464288 0 0 148 1131 641 1053 0 1 99 0 0 0 0 102780 64028 30108 464572 0 0 304 18380 7232 1073 0 1 92 2 5 0 0 102780 63796 30120 464748 0 0 91 23079 8133 1084 1 1 96 0 3 0 0 102780 63780 30132 464844 0 0 189 23260 7857 1165 0 1 93 0 4 0 0 102780 63284 30144 465100 0 0 192 23812 7961 1075 1 2 92 0 6 5 0 102780 62912 30156 465352 0 0 225 14088 5662 1097 0 1 96 0 3 -> Cancelled Transfer (About here) 0 0 102780 62680 30160 465556 0 0 167 111 584 975 0 0 98 0 1 0 0 102780 62416 30180 465652 0 0 140 515 701 946 0 0 98 1 1 So do I interpret that the bottle neck here is CPU? And that since before xen I could get 60MB/s transfer now 8-20MB/s it''s the virtual bridge eating up my CPU. So what would my options be? New quad core CPU? More network cards? Is there a network device out there which can handle the virtual bridge rather than putting the load on the CPU? Any thoughts/suggestions? I purchased a 1Gbit router few months ago I''d like to get my money''s worth out of it. Hobbling along at 8MB/s average is almost what I used to have on 100baseT. Thanks, Michael -- *G. Michael Carter* *Projects* .. GarageSales http://www.garagesaleorganizer.ca Blog/Diaries http://theshadow.carterfamily.ca Family Pictures http://photoalbum.carterfamily.ca GeoCaching http://www.geocache.carterfamily.ca *Contact Info* Office 1-905-890-8100 x 28494 Work Site http://www.oracle.com 1-905-267-8494 Personal http://www.carterfamily.ca Cell 519-215-1869 AIM MikeyCarter Fax 1-519-941-0009 OIM Michael.Carter _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users