I have Dedicated Server on 1and1 with
RAM - 16 GB HD 1000GB CPU: Intel(R) Xeon(R) CPU E3-1270 v3 @ 3.50GHz, 4 cores Centos 6 Apache/2.2.15 PHP 5.3.3 Mysql
I have a video sharing website (on joomla) with about 5000 media items
3500 media items on Amazon and rest is on server (about 489GB) and recently website slowed down significantly. THe pages started to load for 5 – 10 seconds, backend about 10 – 20 seconds. 2 – 3 days before it worked much better, not perfect but much better.I need to identify the problem, myabe you can suggest some commands for tracking performance wich would help me to find the problem?? Because something really went wrong with server.
———-‘top’ result————–
top - 10:35:45 up 17:57, 1 user, load average: 41.21, 46.46, 55.28 Tasks: 327 total, 7 running, 320 sleeping, 0 stopped, 0 zombie Cpu(s): 62.6%us, 0.0%sy, 0.0%ni, 28.8%id, 8.6%wa, 0.0%hi, 0.0%si, 0.0%st Mem: 16294080k total, 15210716k used, 1083364k free, 55924k buffers Swap: 4194288k total, 30268k used, 4164020k free, 11651236k cached 1457 apache 20 0 1167m 157m 4632 S 34.2 1.0 16:17.50 ffmpeg 614 root 20 0 265m 56m 4260 R 33.2 0.4 20:00.14 ffmpeg 1059 root 20 0 265m 55m 4260 R 33.2 0.3 16:52.33 ffmpeg 2733 apache 20 0 266m 55m 4256 R 33.2 0.4 5:25.97 ffmpeg 2786 apache 20 0 266m 56m 4264 R 33.2 0.4 4:50.05 ffmpeg 2854 root 20 0 147m 25m 3580 R 33.2 0.2 3:25.02 ffmpeg 29788 root 20 0 267m 56m 4268 R 33.2 0.4 34:31.32 ffmpeg 31612 root 20 0 266m 55m 4264 R 33.2 0.4 27:22.21 ffmpeg 32589 root 20 0 266m 57m 4264 R 33.2 0.4 24:25.37 ffmpeg 2625 apache 20 0 265m 54m 4256 R 32.9 0.3 6:18.87 ffmpeg 2895 apache 20 0 266m 57m 4260 R 32.9 0.4 2:22.71 ffmpeg 28797 root 20 0 267m 58m 4268 R 32.9 0.4 46:03.94 ffmpeg 32297 root 20 0 266m 57m 4264 R 32.9 0.4 25:24.76 ffmpeg 28343 root 20 0 268m 57m 4268 R 32.6 0.4 54:09.69 ffmpeg 1436 root 20 0 265m 55m 4260 R 32.2 0.4 15:02.93 ffmpeg 2138 root 20 0 265m 55m 4260 R 32.2 0.3 10:41.26 ffmpeg 30435 root 20 0 266m 56m 4264 R 32.2 0.4 30:41.72 ffmpeg 2647 apache 20 0 266m 57m 4272 R 31.9 0.4 6:15.76 ffmpeg 2978 root 20 0 144m 22m 3572 R 31.9 0.1 1:15.28 ffmpeg [root@u17669867 ~]# iostat Linux 2.6.32-431.20.5.el6.x86_64 (u17669867) 08/15/2014 _x86_64_ (8 CPU) avg-cpu: %user %nice %system %iowait %steal %idle 78.37 15.62 0.38 2.05 0.00 3.58 Device: tps Blk_read/s Blk_wrtn/s Blk_read Blk_wrtn sdc 8.57 495.09 2368.59 33674022 161101560 sda 8.41 30.21 171.06 2054642 11634608 sdd 6.85 390.15 2366.70 26536668 160972704 sdb 4.84 2.84 135.32 193088 9203664 md3 4.35 20.51 27.92 1395238 1898808 md1 5.24 2.25 41.40 153236 2816112 md11 303.70 885.20 2368.26 60207426 161079040 dm-0 303.66 885.19 2368.26 60206658 161079048 dm-1 0.12 4.33 0.01 294802 720 dm-2 3.70 16.15 27.91 1098146 1898064 dm-3 0.00 0.02 0.00 1602 24
Answer
Start by tracing what is taking too long with Chrome/IE developer tools (network tab).
Confirming that the slowness is happening for all requests irrespective of type, trace the path from end user to your server (traceroute
) looking for any excessive delays. Test from multiple places.
On the server, check the number of open connections (netstat -antp | grep httpd
), you may be hitting a limit of concurrent users and the connections/requests get queued up waiting for a slot.
Also check CPU (top
, ps aux
), memory (free -m
) and disk usage (iostat
).
EDIT: Your top results show CPU starvation and the iostat results aren’t any better. Consider adding another server or move the CPU-intensive work away from the web server.
Attribution
Source : Link , Question Author : user1584043 , Answer Author : Giovanni Tirloni