Differences

This shows you the differences between two versions of the page.

--- cluster:149 [2016/08/31 14:42] – [Supermicro] hmeij07
+++ cluster:149 [2016/12/06 20:13] (current) – [Supermicro] hmeij07
@@ Line 6: / Line 6: @@
-==== The problem ====
+==== The Storage Problem ====
 In a commodity HPC setup deploying plain NFS, bottle necks can develop.  Then the compute nodes hang and a cold reboot of the entire HPCC is needed. NFS clients on a compute node may contact NFS daemons on our file server sharptail and ask for say a file. The NFS daemon assigned the task then locates the content via metadata (location, inodes, access, etc) on the local disk array.  The NFS daemon  collects the content and hands it off to the NFS client on compute node. So the data passes thru the entire NFS layer.
@@ Line 104: / Line 104: @@
   * Grab the 32x2T flexstorage hard drives and insert into cottontail's empty disk array
       * Makes for a 60T raw raid 6 storage place (2 hot spares)
-      * move the greentail /home backup to it (ready to serve out if needed)
+      * move the sharptail /snapshots to it (remove that traffic from file server)
-      * move the sharptail /snashots to it (remove that traffic from file server)
   * Dedicate greentail's disk array to /sanscratch
       * Remove /home_backup 10T
       * Extend /sanscratch form 27T to 37T
-  * Dedicate sharpttail's disk array to /home
+  * Dedicate sharptail's disk array to /home
       * Keep old 5T /sanscratch as backup, idle
       * Remove 15T /snapshots
       * Extend /home for 10T to 25T
-      * Raise default quota to 512G
+      * Keep 7T /archives until those users graduate, move to Rstore
-      * Keep 7T /archives until those users graduate, move to Rstore (add to old /sanscratch?)
 **Long term plan**
   * Start a BeeGFS storage cluster
-      * n38-n45 (8) metadata servers (15K local disk, no raid)
+      * cottontail as MS (management server)
-      * Buy 2x 2U Supermicro storage server for a total of 80T usable, raid 6 ($14K)
+      * sharptail as AdMon (monitor server) and proof of concept storage OSS
+          * pilot storage on  idle /sanscratch/beegfs/
+          * also a folder on cottonttail:/snapshots/beegfs/
+      * n38-n45 (8) as MDS (metadata servers, 15K local disk, no raid)
+      * Buy 2x 2U Supermicro for OSS (object storage servers for a total of 80T usable, raid 6, $12.5K)
       * Serve up BeeGFS file system using IPoIB
       * Move /home to it
@@ Line 131: / Line 133: @@
   * backup Openlava scheduler for cottontail.wesleyan.edu
   * backup replacement for greentail.wesleyan.edu (in case it fails)
+Bought.
+ --- //[[hmeij@wesleyan.edu|Henk]] 2016/10/27 14:49//
 Warewulf golden image it as if it is greentail.