thePROBLEMS with theHIVE


SYMPTOM:
Could not mount all 64 home directories on each machine.
PROBLEM:
Number of mountable file systems was set to 64.

SOLUTION:
- set NR_SUPER to 256 in file /usr/src/linux/include/linux/fs.h
"#define NR_SUPER 256"

SYMPTOM:
Could not list (ls) all the mount points (64) in the directory where they were mounted (/homes).
PROBLEM:
Read directory cache was to small.

SOLUTION:
- set NFS_READDIR_CACHE_SIZE to 256 in file /usr/src/linux/include/linux/nfs_fs.h
"#define NFS_READDIR_CACHE_SIZE 256"
- note that this is the largest value that NFS_READDIR_CACHE_SIZE can be set to since it must fit within a single 4K byte page.

SYMPTOM:
As many as 1 in 1000 'rsh' commands hang. This gets worse with contention on the network.
PROBLEM:
- Unknown

SOLUTION:
- none yet

SYMPTOM:
The DRONE (D0) runs out of processes, which are necessary to start tasks on the bees.
PROBLEM:
Since there are 64 BEEs, the DRONE's process table can fill up quickly. This gets worse when multiple users are running on theHIVE.

SOLUTION:
- none yet

John E. Dorband / NASA GSFC / John.E.Dorband@gsfc.nasa.gov