2011年10月14日 8:44Hi All,
Thanks in advance for your help. I am using HPC R2 pack installed on 2008 R2 machine acting as head node, I had compute node HPCCN01-07 added to cluster & everything working fine. Because of some hardware failure I was in need to replace two of my nodes HPCCN03 & HPCCN07. I deleted both nodes from HPC server manager GUI & operation went successful. NOw I am trying to add two different machine with same hostname to the headnode as preconfigured compute nodes, the operation is failing with following errors in provisioning logs :
Failed to update the scheduler configuration. The node with name HPCCNO7 already exists.
Any ideas how to get past this errors.
2011年10月14日 14:58Did you recreate computer account on Active Directory? I am not sure, but try to delete computer accounts for HPCCN03 & HPCCN07 from AD, and rejoin this nodes to domain firstly, then add it to cluster.
2011年10月15日 4:16Already tried but no help. However on head node I reinstalled the cluster manager software and now I am able to add all the nodes. I believe the cluster manager keeps a backup of information somewhere and when it is trying to update that it failed. It would be good to know about such details. Otherwise whenever a problem occurs always reboot or reinstallation is not feasible.
Yes, thats a problem. HPC pack tracks information about its state and all of nodes. Sometimes, for example, as in your situation, it can produce an error. Sad, but i didnt find a solution for such a problems. But i'll ask some people who may know. Because reinstalling of whole bunch of stuff when such an error occurs, is no good..
BTW, this error does not connect with AD-infrastructure. I think that it keeps its database in another place.
HPC Pack has enough not-so-well-documented-features. :(
The HPC maintains several SQL databases. We really need to have documentation for digging those databases or workarounds for situation like mine without re-installation or reboot :).
- 已编辑 Rohit_k13 2011年10月17日 10:03