PHP failed to ptrace(PEEKDATA) pid 1918: Input/output error (5)或者 increase pm.start_servers解決

今天網站出現負載很大，cpu占用率超高，網站延遲很大，前臺出現此錯誤信息：

[19-Sep-2016 23:30:38] WARNING: [pool www] child 31954, script '/var/www/html/casino.21pink.com/public/index.php' (request: "POST /index.php") executing too slow (2.407095 sec), logging
[19-Sep-2016 23:30:38] WARNING: [pool www] child 31943, script '/var/www/html/casino.21pink.com/public/index.php' (request: "POST /index.php") executing too slow (2.062648 sec), logging
[19-Sep-2016 23:30:38] WARNING: [pool www] child 31941, script '/var/www/html/casino.21pink.com/public/index.php' (request: "POST /index.php") executing too slow (2.666022 sec), logging
[19-Sep-2016 23:30:38] WARNING: [pool www] child 30382, script '/var/www/html/casino.21pink.com/public/index.php' (request: "POST /index.php") executing too slow (2.021010 sec), logging
[19-Sep-2016 23:30:38] WARNING: [pool www] child 29797, script '/var/www/html/casino.21pink.com/public/index.php' (request: "POST /index.php") executing too slow (2.554333 sec), logging
[19-Sep-2016 23:39:27] WARNING: [pool www] seems busy (you may need to increase pm.start_servers, or pm.min/max_spare_servers), spawning 8 children, there are 2 idle, and 68 total children
[19-Sep-2016 23:40:58] WARNING: [pool www] seems busy (you may need to increase pm.start_servers, or pm.min/max_spare_servers), spawning 8 children, there are 0 idle, and 81 total children
[19-Sep-2016 23:40:59] WARNING: [pool www] seems busy (you may need to increase pm.start_servers, or pm.min/max_spare_servers), spawning 16 children, there are 3 idle, and 86 total children
[19-Sep-2016 23:41:00] WARNING: [pool www] seems busy (you may need to increase pm.start_servers, or pm.min/max_spare_servers), spawning 32 children, there are 2 idle, and 88 total children
[19-Sep-2016 23:41:01] WARNING: [pool www] seems busy (you may need to increase pm.start_servers, or pm.min/max_spare_servers), spawning 32 children, there are 0 idle, and 91 total children

從信息看，應該是php進程較少：

經過查看發現進程確實很少，解決如下：

修改配置文件增加進程

vim /usr/local/php/etc/php-fpm.ini
pm = dynamic  #動態模式
pm.max_children = 512  #靜態開啟的進程數
pm.start_servers = 128 #動態模式的起始進程數
pm.min_spare_servers = 128 #動態模式最小進程數
pm.max_spare_servers = 512 #動態模式最大進程數

數值設置，參考自己的實際硬件配置，可以參考 內存/20M 來計算。

重新加載進程

/etc/init.d/php-fpm reload

等待了一點時間發現又報錯：

[19-Sep-2016 23:30:20] WARNING: [pool www] child 7782, script '/var/www/html/casino.21pink.com/public/index.php' (request: "POST /index.php") executing too slow (2.224150 sec), logging
[19-Sep-2016 23:30:08] ERROR: failed to ptrace(PEEKDATA) pid 7663: Input/output error (5)
[19-Sep-2016 23:30:20] WARNING: [pool www] child 31952, script '/var/www/html/casino.21pink.com/public/index.php' (request: "POST /index.php") executing too slow (2.045803 sec), logging
[19-Sep-2016 23:30:20] ERROR: failed to ptrace(PEEKDATA) pid 31952: Input/output error (5)

查找資料如下：

It usrears you have request_slowlog_timeout enabled. This normally takes any request longer than N seconds, logs that it was taking a long time, then logs a stack trace of the script so you can see what it was doing that was taking so long.

In your case, the stack trace (to determine what the script is doing) is failing. If you're running out of processes, it is because either:

After php-fpm stops the process to trace it, the process fails to resume because of the error tracing it

The process is resuming but continues to run forever.

My first guess would be to disable request_slowlog_timeout. Since it's not working right, it may be doing more harm than good. If this doesn't fix the issue of running out of processes, then set the php.ini max_execution_time to something that will kill the script for sure.

大概說是設置了slowlog和php的最大執行時間。

解決：

禁止slowlog

vim php-fpm.conf
;request_slowlog_timeout = 10s
;slowlog = /usr/local/log/php-fpm/ckl-slow.log

修改最大執行時間：

vim php.ini
max_execution_time = 60

重啟進程：

/etc/init.d/php-fpm reload

等待一段時間，發現一切正常。

查看TCP連接相關：

[root@GreenZonewebapp2 php-fpm.d]# ss -s
Total: 323 (kernel 675)
TCP:   4538 (estab 12, closed 4518, orphaned 0, synrecv 0, timewait 4518/0), ports 0
Transport Total     IP        IPv6
*      675       -         -        
RAW      0         0         0        
UDP      5         4         1        
TCP      20        19        1        
INET      25        23        2        
FRAG      0         0         0

同時發現系統TIMEWAIT 較多，所以優化了一些內核相關參數 

#vim /etc/sysctl.conf   進行內核的相關優化
# sysct -p
bash: sysct: command not found
[root@sapi etc]# sysctl -p
net.ipv4.ip_forward = 0
net.ipv4.conf.default.rp_filter = 1
net.ipv4.conf.default.accept_source_route = 0
net.ipv4.tcp_max_tw_buckets = 6000
net.ipv4.ip_local_port_range = 1024 65000
net.ipv4.tcp_tw_recycle = 1
net.ipv4.tcp_tw_reuse = 1
net.ipv4.tcp_syncookies = 1
net.ipv4.tcp_max_orphans = 262144
net.ipv4.tcp_keepalive_time = 1200
net.ipv4.tcp_keepalive_intvl = 30
net.ipv4.tcp_keepalive_probes = 3
net.ipv4.tcp_synack_retries = 1
net.ipv4.tcp_syn_retries = 2
net.ipv4.tcp_max_orphans = 262144
net.ipv4.tcp_max_syn_backlog = 262144
net.ipv4.tcp_timestamps = 0
net.core.rmem_max = 16777216
net.core.wmem_max = 16777216
net.ipv4.tcp_rmem = 4096 87380 4194304
net.ipv4.tcp_wmem = 4096 16384 4194304
net.core.wmem_default = 8388608
net.core.rmem_default = 8388608
net.core.rmem_max = 16777216
net.core.wmem_max = 16777216
net.core.somaxconn = 262144
kernel.sysrq = 0
kernel.core_uses_pid = 1
kernel.msgmnb = 65536
kernel.msgmax = 65536
kernel.shmmax = 68719476736
kernel.shmall = 4294967296
vm.swappiness = 0
fs.file-max = 409600

  過一陣再查看：

# ss -s
Total: 281 (kernel 362)
TCP:   520 (estab 22, closed 493, orphaned 0, synrecv 0, timewait 493/0), ports 475
 
Transport Total     IP        IPv6
*         362       -         -        
RAW       0         0         0        
UDP       1         1         0        
TCP       27        27        0        
INET      28        28        0        
FRAG      0         0         0