<div class="MsoNormal" style="">Dear all,<br> The same problem is occurred, and I cannot kill the following crashed jobs.<br> qdel 41074<br> qdel: Server could not connect to MOM 41074.h101.cl.unh.edu<br> qdel 41075<br> qdel: Server could not connect to MOM 41075.h101.cl.unh.edu<o:p></o:p></div> <div class="MsoNormal" style="">The last one was killed by Tod, and I appreciate him. But, it seems that this is a frequently occurred problem. What can we do to avoid putting in trouble the administrators frequently to kill our crashed jobs? If for any reasons some nodes running a job fail, e.g., 2 nodes out of 5 requested nodes fail, we cannot do qdel the job. Thus in such a case the other nodes, e.g., here in the later example 3 nodes, are still involved for solving a crashed job. In this case it should be existed a way to make free the involved nodes. Unfortunately, qdel command does not work in this case, and I appreciate if you let the users of Zaphod
know another way to get ride of such a sever problem. <br> </div> <div class="MsoNormal" style="">Thank you in advance for your contributions,<br> Your,<br> S. Jalali.<o:p></o:p></div> <p> 
<hr size=1>Don't be flakey. <a href="http://us.rd.yahoo.com/evt=43909/*http://mobile.yahoo.com/mail">Get Yahoo! Mail for Mobile</a> and <br><a href="http://us.rd.yahoo.com/evt=43909/*http://mobile.yahoo.com/mail">always stay connected</a> to friends.