Pending AIO operation, recycle later.... Lot of Errors

[sahith]

I am getting lot of errors like

[164.58.218.66:64434-0] Pending AIO operation, recycle later...
[164.58.218.66:64434-0] Pending AIO finished, recycle connection.

My error log is almost full of these errors. I need help over clearing these errors please.

System Config :
Dual Intel(R) Xeon(R) CPU E5-2620 0 @ 2.00GHz
64GB of RAM (only 8% being utilized at present)
2x300GB SAS Drivers in RAID1
1GBPS WAN.

 

[mistwang]

those are not errors, just the way AIO works.
what is the I/O wait of your server in the top output? show us the top output
You need to find out the root cause of the high I/O wait and minimized it. since you have 64GB memory, you can move PHP session, and other temp files to /dev/shm instead of /tmp.

 

[sahith]

May i know how to move tmp files to/dev/shm. And which top command output you want

 

[sahith]

top - 23:54:09 up 89 days, 16:41, 2 users, load average: 3.33, 3.87, 21.62
Tasks: 536 total, 4 running, 530 sleeping, 2 stopped, 0 zombie
Cpu(s): 9.1%us, 2.9%sy, 0.0%ni, 86.0%id, 1.0%wa, 0.0%hi, 1.0%si, 0.0%st
Mem: 65944680k total, 41812312k used, 24132368k free, 257488k buffers
Swap: 2097144k total, 44k used, 2097100k free, 36174768k cached

 

[mistwang]

edit php.ini, change session save path to /dev/shm

 

[mistwang]

If use LiteSpeed page cache, change page cache storage path to /dev/shm/lscache

 

[sahith]

How much space allocated to /dev/shm. Is it RAMDISK?

 

[mistwang]

Yes, it is ramdis,by default linux give half of total memory to /dev/shm, you server should have 32GB. You can change it from /etc/fstab.

 

[sahith]

Yes, it is ramdis,by default linux give half of total memory to /dev/shm, you server should have 32GB. You can change it from /etc/fstab.

THank you for your attention. You are so much helpful wherever i was needed, i will implement those recommendations at first.

 

[sahith]

Implementing this go so many problems later, now litespeed is not even working. Even after force reinstall. I need help please.

 

[NiteWave]

check of /dev/shm is full
create a cronjob to clean up old files on /dev/shm to ensure it's not full
also run
df -hi
to see if inode is full

 

[sahith]

No i reverted back to /tmp, to make sure it works fine. Anyway i even disabled litespeed cache. Whatever litespeed is not working. Even if switched to litespeed.

• Switched to LiteSpeed Successfully
• Apache Stopped.
• LiteSpeed started successfully.
• Apache port offset has been set to 0.

From then i am getting 404 error pages for all sites displaying "
Index of /"

 

[sahith]

I even getting the follwing error sometimes when switching back to Apache,

root@server [~]# service httpd restart
httpd not running, trying to start
(98)Address already in use: AH00072: make_sock: could not bind to address [::]:8 0
(98)Address already in use: AH00072: make_sock: could not bind to address 0.0.0. 0:80
no listening sockets available, shutting down
AH00015: Unable to open logs

 

[NiteWave]

if it's WHM/cPanel, please go
WHM->Litespeed Plugin
to switch between apache and litespeed

is the 404 error fixed ?

 

[sahith]

I hired a server admin, who fixed it. It was caused due some permissions issue.