Both the internal drive and the <br>one hanging off the USB 2 cable<br>are IDE.<br><br>When I run the backup, the <br>machine is doing little else, so I'd<br>guess the CPU is not particularly<br>loaded.<br><br>I'm a little puzzled as to why the
<br>hdparm "d" parameter does not<br>function, but the write says hdparm<br>does not work with all chip sets. <br><br>Well, for now the backups work.<br>I just have to wait. The transfer<br>rate is the 8 Mbps I mentioned. Yet
<br>on another USB 2 port I can move<br>data between two machines over <br>a wireless link at 45 to 50 Mbps.<br><br>So I think there is a mystery here<br>yet. I'd like to see at least that when<br>backing up to an external drive on
<br>a USB 2 port.<br><br>Someday...<br><br> -Bob<br><br><p><DEFANGED_div class="gmail_quote"><blockquote class="gmail_quote" DEFANGED_style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"><p><DEFANGED_div><p><DEFANGED_div class="Wj3C7c">
<blockquote class="gmail_quote"><p><br>hdparm -cuda /dev/x</p></blockquote></p><DEFANGED_div></p><DEFANGED_div><p><br>I forgot about one thing Bob.<br>hdparm only works for IDE drives, since your drive is interfaced via USB it must be /dev/s-something right?
<br><br>if that it the case than DMA shouldn't be needed at all.<br><br>As the purpose of DMA is to free up the processor maybe you should watch the Load Avg when copying a file to see if it's<br>abnormally high.
<br>
<br></p><br><blockquote class="gmail_quote"><p><DEFANGED_div class="Ih2E3d"><p>only reports for a: "readahead = 256 (on) <br><br>Running "hdparm -d1 /dev/x"
<br>result in "HDIO_SET_DEM failed: Invalid argument."<br><br>
The hdparm man page suggest that<br>in rare cases one has to use the -X<br>parameter. It also says this can cause<br>real headaches.<br><br> -Bob<br><br></p><p></p><br>
<br></p><DEFANGED_div></blockquote></blockquote></p><DEFANGED_div><br>