ORA-2730*引出的系統參數vm.min_free_kbytes理解

oerr ORA-2730*

[oracle@p19c01 ~]$ oerr ORA 27300
27300, 00000, "OS system dependent operation:%s failed with status: %s"
// *Cause:  OS system call error
// *Action: contact Oracle Support
[oracle@p19c01 ~]$ oerr ORA 27301
27301, 00000, "OS failure message: %s"
// *Cause:  OS system call error
// *Action: contact Oracle Support
//
[oracle@p19c01 ~]$ oerr ORA 27302
27302, 00000, "failure occurred at: %s"
// *Cause:  OS system call error
// *Action: contact Oracle Support
//

*Cause:  OS system call error

 The errors are usually caused by OS system call error or OS configuration issue 。

能夠發現，若是數據庫trace日誌中出現ORA-2730*的錯誤，很大多是因爲OS system的錯誤致使。

 

案例分析：

Tue Sep 01 04:05:33 2020
skgxpvfynet: mtype: 61 process 417356 failed because of a resource problem in the OS. The OS has most likely run out of buffers (rval: 4)
Errors in file /u01/app/oracle/diag/rdbms/*****/******/trace/******1_w001_417356.trc (incident=96021):
ORA-00603: ORACLE server session terminated by fatal error
ORA-27504: IPC error creating OSD context
ORA-27300: OS system dependent operation:sendmsg failed with status: 105
ORA-27301: OS failure message: No buffer space available
ORA-27302: failure occurred at: sskgxpsnd2
Incident details in: /u01/app/oracle/diag/rdbms/*****/******/incident/incdir_96021/******1_w001_417356_i96021.trc
opidrv aborting process W001 ospid (417356) as a result of ORA-603

以上爲某一套數據庫trace日誌中報錯的信息，能夠發現幾個關鍵點：

一、status: 105

二、 failure occurred at: sskgxpsnd2

三、No buffer space available。

Troubleshooting ORA-27300 ORA-27301 ORA-27302 Errors (Doc ID 579365.1)    

經過MOS文章能夠查詢到與上述報錯相符的信息：

See: Oracle Linux: ORA-27301:OS Failure Message: No Buffer Space Available (Doc ID 2041723.1)

關於解決方案中提到的vm.min_free_kbytes參數，單位是KB，Oracle建議的取值範圍是0.4%到5%之間，若是內存<=32G，能夠忽略不設置。

查看系統vm.min_free_kbytes默認設置大小：

[root@p19c01 ~]# sysctl -a | grep vm.min_free_kbytes
vm.min_free_kbytes = 11433

查看系統開啓NUMA的node數量：

numactl --hardware | grep available: | awk '{print $2}'

Notes：該參數表示Linux VM最低保留多少的空閒內存空間，當可用的內存低於配置參數時，系統會進行cache內存的回收，來進行內存的釋放。

★ 參數簡述
該Linux系統參數的功能是用來設置一個最小內存空間給系統內核使用
該值設置過大會浪費空間，保留太小會形成系統壓力
★ 知識點
※ 用途：合理設置該值有助於Linux系統更有效地回收內存
※ 注意：對於內存較小（≤32GB）的環境不建議設置，保持默認便可
※ 危險：不要對正在運行的環境進行設置
※ 有修改操做，非原版
※ vm.min_free_kbytes參數的單位是：KB
※ 內存水位
    min水位：下的內存是保留給內核使用的；當到達min，會觸發內存的direct reclaim
    low水位：比min高一些，當內存可用量小於low的時候，會觸發 kswapd回收內存
    high水位：繼續睡眠
※ 內存回收方式
    direct reclaim : 觸發min水位線時執行
    kswapd reclaim : 觸發low水位線時執行

這裏摘自其餘博主的一個計算腳本能夠用來計算vm.min_free_kbytes，直接複製到shell控制檯執行便可：

MIN_FREE_KBYTES_SYSCTL=$(egrep ^vm.min_free_kbytes /etc/sysctl.conf | awk '{print $3}')
MIN_FREE_KBYTES_MEMORY=$(cat /proc/sys/vm/min_free_kbytes)
NUMA_NODE_COUNT=$(numactl --hardware | grep available: | awk '{print $2}')
TOTAL_MEMORY_KBYTES=$(free -k | awk '/Mem:/ {print $2}')
NUMA_BASED=$(( $NUMA_NODE_COUNT * 1048576 ))
MEMORY_BASED=$(( $TOTAL_MEMORY_KBYTES / 200 ))
if [[ $NUMA_BASED -ge $MEMORY_BASED ]]
then
	RECOMMEND_VALUE=$NUMA_BASED
else
	RECOMMEND_VALUE=$MEMORY_BASED
fi
OFFSET=$(echo $RECOMMEND_VALUE*.05 | bc | cut -d"." -f1)
LOWER_BOUND=$(echo $RECOMMEND_VALUE-$OFFSET | bc)
UPPER_BOUND=$(echo $RECOMMEND_VALUE+$OFFSET | bc)
if [[ $MIN_FREE_KBYTES_SYSCTL -ge LOWER_BOUND && $MIN_FREE_KBYTES_SYSCTL -le UPPER_BOUND ]]
then
	SYSCTL_IN_RANGE=YES
else
	SYSCTL_IN_RANGE=NO
fi
#sysctl in range?
if [[ $MIN_FREE_KBYTES_MEMORY -ge LOWER_BOUND && $MIN_FREE_KBYTES_MEMORY -le UPPER_BOUND ]]
then
	MEMORY_IN_RANGE=YES
else
	MEMORY_IN_RANGE=NO
fi
DETAIL=$(
echo -e "Total Memory:       $TOTAL_MEMORY_KBYTES";
echo -e "NUMA node count:    $NUMA_NODE_COUNT";
echo -e "NUMA calculated:    $NUMA_BASED";
echo -e "memory calculated:  $MEMORY_BASED";
echo -e "recommended value:  $RECOMMEND_VALUE";
echo -e "permitted range:    $LOWER_BOUND to $UPPER_BOUND";
echo -e "in sysctl.conf:     $MIN_FREE_KBYTES_SYSCTL";
echo -e "sysctl in range?:   $SYSCTL_IN_RANGE";
echo -e "in active memory:   $MIN_FREE_KBYTES_MEMORY";
echo -e "memory in range?:   $MEMORY_IN_RANGE";
)
if [[ $SYSCTL_IN_RANGE = YES && $MEMORY_IN_RANGE = YES ]]
then
	echo -e "SUCCESS: vm.min_free_kbytes is configured as recommended.  Details:\n\n$DETAIL"
elif [[ $MIN_FREE_KBYTES_SYSCTL -lt $LOWER_BOUND || $MIN_FREE_KBYTES_MEMORY -lt $LOWER_BOUND ]]
then
	echo -e ":: Result : 【FAILURE】: vm.min_free_kbytes is not configured as recommended"
	ZZT_V1=$(( $NUMA_NODE_COUNT * 1 * 1024 * 1024 ))
	ZZT_V2=$(( $TOTAL_MEMORY_KBYTES * 5 /10 / 100 ))
	if [ $ZZT_V1 -gt $ZZT_V2 ]
	then
	ZZT_MAX=$ZZT_V1
	else
	ZZT_MAX=$ZZT_V2
	fi
	ZZT_MAX_GB=$(($ZZT_MAX/1024/1024))
	echo ">>> if output is <FAILURE>，please vi sysctl.conf and edit param's value for:vm.min_free_kbytes and reboot"
	echo ">>> [formula_oracle]vm.min_free_kbytes value (Kb) =MAX(1GB * number_numa_nodes, 0.5% * total_memory) "
	echo ">>> [calculated_zzt]vm.min_free_kbytes = $ZZT_MAX   (About: $ZZT_MAX_GB GB)"
	ZZT_V3=32
	ZZT_V4=$(($TOTAL_MEMORY_KBYTES/1024/1024))
	if [ $ZZT_V4 -gt $ZZT_V3 ]
	then
		echo ">>> [PASS]The current memory is suitable for setting system parameters."
	else
		echo ">>> [WARN]Your memory is too small to set this parameter."
	fi
	echo -e "Details:\n\n$DETAIL"
elif  [[ $MIN_FREE_KBYTES_SYSCTL -gt $UPPER_BOUND && $MIN_FREE_KBYTES_MEMORY -gt $UPPER_BOUND ]]
then
  echo -e "WARNING: vm.min_free_kbytes is not configured as recommended.  Details:\n\n$DETAIL"
else
  echo -e "ERROR: Inconsistent results.  Details:\n\n$DETAILS"
fi

測試結果以下：