System completely freezes and becomes unresponsive

Question

Hi,
We have a single processor, 1 GB RAM Redhat system (desktop). Normally with the database(oracle 10.1g) and client connections running, the memory utilization of this server is around 90-99% . However every few days the box completely freezes and becomes unresponsive (users cannot login, client connections fail). We have to hardboot the box and restart the database. When in stuck state, the log files are not updated, except that sendmail logs heavy load average. ( Messages like "Load Avg 12: 1 Times(s) Load Avg 13: 1 Times(s) Load Avg 14: 1 Times" and so on) and the free memory falls to 8 mb or less but with above 2 gb swap space available.
My question is if there is no free memory, system should thrash (excessive swapping) and become slow, but instead our box completely freezes, though the interactive response ( for any connection which is still active) and ping response is ok, they aren't slow.
Is there any kernel level parameter that can be tuned so the server won't freeze, (tries to move the memory pages to swap space or any other way). When the server is stuck (freezes ) any ssh connections still active will stay on, but new connections fail. One time, using that session, I tried to free up memory by killing processes like X Server etc, but still the server didn't unfreeze.
Server info:
Red Hat Enterprise Linux AS release 3 (Taroon Update 4)
Kernel (uname -r) 2.4.21-27.EL
Intel Pentium 4 CPU 2.80GHz (Single processor)
1 GB RAM
Oracle db version - 10.1.0.3.0
Lastly, I don't think this is any hw problem, because we see similiar problem on multiple system with similiar configs.
Please guide how to resolve the problem. Thanks.

Answer 1

Hi Praveen, thanks for your inputs.
Andrey - We've 3 log files. rest info -
SQL> select * from v$log;
GROUP# THREAD# SEQUENCE# BYTES MEMBERS ARC STATUS FIRST_CHANGE# FIRST_TIM
1 1 5599 10485760 1 NO CURRENT 561614475 25-NOV-05
2 1 5597 10485760 1 NO INACTIVE 561603206 25-NOV-05
3 1 5598 10485760 1 NO ACTIVE 561608824 25-NOV-05
SQL> select * from v$instance_recovery;
RECOVERY_ESTIMATED_IOS ACTUAL_REDO_BLKS TARGET_REDO_BLKS LOG_FILE_SIZE_REDO_BLKS LOG_CHKPT_TIMEOUT_REDO_BLKS LOG_CHKPT_INTERVAL_REDO_BLKS FAST_START_IO_TARGET_REDO_BLKS TARGET_MTTR ESTIMATED_MTTR CKPT_BLOCK_WRITES OPTIMAL_LOGFILE_SIZE ESTD_CLUSTER_AVAILABLE_TIME WRITES_MTTR WRITES_LOGFILE_SIZE WRITES_LOG_CHECKPOINT_SETTINGS WRITES_OTHER_SETTINGS WRITES_AUTOTUNE WRITES_FULL_THREAD_CKPT
70 501 5305 18432 5305 0 12 35578 0 139734 0 0 6467 697