From: Vimal Upreti (Vimal@ohitelecom.com)
Date: Sat Nov 30 2002 - 01:30:18 EST
Hi all,
We are in the process of configuring Oracle 9i Release 2 on TRU64 Cluster (2 ES40s with RA3000 ). So far we have done just the installation( No database creation).
The problem is that System looks very slow. Performance is perfectly OK when we perform some activity on single node, but the moment we try to do same thing on both the nodes, it takes hell lot of time. Even while running root.sh this was very slow on both the nodes (almost hung- took at least 10 good minutes). On the Console Log following messages were displayed continuously:
SCSI EVENT [600]
SCSI EVENT [600]
CAAD [1049124]:RTD #0: Action script /var/cluster/caa/script/cluster_locked.scr
CAAD [1049124]:RTD #0:An error was encountered while polling 'Cluster_locked' (Check) timeout!
SCSI EVENT [700]: SCSI error occured on Bus2 Target7 Lun7
....................Bus2 Target7 Lun7 is directing to one of the Diff SCSI Controller.
While starting gsd domain, from one node it takes fraction of second, while from other it takes atleast 10-15 seconds.
Is IT Normal???
**************************************************
Following are Kernal Parameters, that I tried to change before installation...
proc:
max_per_proc_address_space = 2 gig
per_proc_address_space = 2gig
per_proc_stack_size = 33554432
max_per_proc_stack_size = 500m
per_proc_data_size = 201326592
vm:
new_wire_method = 0
vm_swap_eager = 1 "Earlier this was set as vm-swap-eager = 1 and not vm_swap_eager =1 "
ubc_maxpercent = 70
gh_chunks = 518 " I tried setting this to 518 but after booting this comes out to be 0 "
vfs
fifo_do_adaptive defaults = 0
ipc:
shm_max = 2139095040
shm_mni = 256
shm_min = 1
shm_seg = 1024
ssm_threshold = 0 "I tried setting this to be 0 but after booting this has become a very large value (atleast 20 digit value)"
rdg:
msg_size = 32768
max_objs = 5120
max_async_req = 256
max_sessions = 200
rdg_max_auto_msg_wires = 0
rdg_auto_msg_wires = 0
rt:
aio_task_max_num = 300 "I tried setting this to be 300 but after booting this has become 307"
***************************************************************
Can you please let me know whether this is a Hardware configuration problem or Kernel Parameter?
Cluster works fine in normal conditions (Individual access of the servers).
Appreciate your help,
Vimal Upreti
This archive was generated by hypermail 2.1.7 : Sat Apr 12 2008 - 10:49:00 EDT