Hello all,
I am running Tru64 5.1B with trucluster 5.1B on an ES80. Memory is 8GB with
eight processors and 8GB swap space.. Before this I was running the same
installation of oracle with a GS60E with the same amount of memory and oracle
had been set to use 7GB of memory. I never had any cases of swapping.
However since the change over to the ES80 even with memory not completely
used up at about 80% it begins to swap and yet swap is set to lazy mode. I
would like to tune the sysconfigtab to avoid swapping completely. Current
values in sysconfigtab are as follows:
vm:
vm_page_free_reserved = 20
vm_page_free_min = 30
vm-swap-eager = 0
ubc_minpercent = 1
ubc_maxpercent = 5
ubc_borrowpercent = 2
vm_page_free_target = 2048
vm_segmentation = 1
swapdevice = /dev/disk/dsk102b
lsm:
lsm_rootdev_is_volume = 0
clubase:
cluster_expected_votes = 3
cluster_name = cauga01
cluster_node_name = sauga01
cluster_node_inter_name = nsauga01-ics0
cluster_node_votes = 1
cluster_interconnect = tcp
cluster_seqdisk_major = 19
cluster_seqdisk_minor = 850
cluster_qdisk_major = 19
cluster_qdisk_minor = 866
cluster_qdisk_votes = 1
sec:
acl_mode = disable
ics_ll_tcp:
ics_tcp_inetaddr0 = 10.1.0.21
ics_tcp_netmask0 = 255.255.255.0
ics_tcp_adapter0 = alt0
vfs:
fifo_do_adaptive = 0
noadd_exec_access = 0
ipc:
shm_max = 2139095040
proc:
max_proc_per_user = 2048
max_threads_per_user = 4096
per_proc_data_size = 8589934592
max_per_proc_data_size = 8589934592
max_per_proc_address_space = 8589934592
per_proc_address_space = 8589934592
maxusers = 8192
exec_disable_arg_limit = 1
max_per_proc_stack_size = 33554431
per_proc_stack_size = 8388609
net:
ifqmaxlen = 65535
inet:
ipport_userreserved = 65535
ipqmaxlen = 65535
tcp_recvspace = 262144
tcp_sendspace = 262144
udp_recvspace = 262144
udp_sendspace = 262144
socket:
sb_max = 16777216
Received on Thu Nov 27 2003 - 05:03:47 NZDT