It's interesting that this problem started at almost exactly midnight Thursday night (Pacific time) and improved significantly at exactly 4 PM today. It's still taking 8-20 seconds for the page generation so there is still an issue.
io_status shows:
OK io_count=850802 io_rate=27 io_tokens=200000 token_refill=512 token_max=200000
vmstat 1 shows:
Code:
procs memory swap io system cpu
r b w swpd free buff cache si so bi bo in cs us sy id
0 0 0 1068 12864 12044 135524 0 0 19 17 138 44 24 41 35
1 0 0 1068 12728 12052 135652 0 0 132 0 242 101 26 35 38
1 0 0 1068 12728 12052 135652 0 0 0 0 106 11 42 58 0
1 0 0 1068 12728 12052 135652 0 0 0 0 105 11 35 65 0
1 0 0 1068 12728 12052 135652 0 0 0 0 105 13 32 68 0
1 0 0 1068 12724 12056 135652 0 0 0 91 195 14 33 67 0
1 0 0 1068 12724 12056 135652 0 0 0 0 106 11 42 58 0
0 0 0 1068 12712 12056 135664 0 0 0 0 107 25 2 6 92
1 0 0 1068 12640 12056 135664 0 0 0 0 108 19 0 5 95
0 0 0 1068 12452 12060 135664 0 0 0 89 198 38 0 0 100
0 0 0 1068 12612 12064 135748 0 0 85 46 240 82 3 5 92
1 0 0 1068 12544 12064 135824 0 0 60 0 177 157 11 30 60
1 0 0 1068 12544 12064 135824 0 0 0 0 105 13 29 71 0
1 0 0 1068 12544 12064 135824 0 0 0 0 103 12 25 75 0
1 0 0 1068 12544 12064 135824 0 0 0 0 103 11 51 49 0
1 0 0 1068 12540 12068 135824 0 0 0 130 234 16 40 60 0
1 0 0 1068 12540 12068 135824 0 0 0 0 104 10 41 59 0
0 0 0 1068 12540 12068 135824 0 0 0 0 107 25 18 20 62
0 0 0 1068 12540 12068 135824 0 0 0 0 102 15 0 0 100
0 0 0 1068 12540 12068 135824 0 0 0 0 102 9 0 0 100
0 0 0 1068 12540 12068 135824 0 0 0 26 128 20 0 0 100
0 0 0 1068 12540 12068 135824 0 0 0 0 102 9 0 0 100
0 0 0 1068 12540 12068 135824 0 0 0 0 102 11 0 0 100
Also, CPU is running at close to 100% with 60%-75% in the system.