Hi,<br><br>It works now.<br><br>Not write (just based on guess) might <br>the md.log step such as is 11140000<br>while use thread I noticed the actually run step started from 11135000<br><br>so I run until it can write to the md.log<br>
<br>then switch to 1 node to run for a while,<br><br>then switch to more nodes.<br><br>But there might be some reason there which I don't know.<br><br>Thanks,<br><br><div class="gmail_quote">On Sun, Sep 18, 2011 at 4:35 PM, lina <span dir="ltr"><<a href="mailto:lina.lastname@gmail.com">lina.lastname@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">Hi,<br><br>Very sporadically and also with high frequent, <br><br>The job I submitted only running without writing (this job is not un-started one, mainly one I stopped and rerun).<br>
<br>Before I thought I did not wait long enough, such as hours, but seriously after 3 or 8 hours, still running not writing.<br>
<br>I ssh to each nodes, all is fully running.<br><br>The storage is NFS, I/O flow can't be choked for hours. <br><br>Really headache, sometimes it works. <br><br>I have had no clue about it. <br><br><br>Thanks for any suggestions,<br>
<font color="#888888">
<br><br>-- <br>Best Regards,<br><br>lina<br><br><br>
</font></blockquote></div><br><br clear="all"><br>-- <br>Best Regards,<br><br>lina<br><br><br>