服务器CPU降频

最近服务器忽然变得很慢，很慢。排查了各种问题，各种怀疑，各种折腾，最后发现是CPU温度过高，降频了，降频了，降频了。
记录一下：
1，内核日志影子

[ 2761.502936] perf: interrupt took too long (10788 > 10748), lowering kernel.perf_event_max_sample_rate to 18000
[ 3775.114536] perf: interrupt took too long (13698 > 13485), lowering kernel.perf_event_max_sample_rate to 14000
[ 4896.214284] perf: interrupt took too long (17275 > 17122), lowering kernel.perf_event_max_sample_rate to 11000
[ 6379.029543] perf: interrupt took too long (25326 > 21593), lowering kernel.perf_event_max_sample_rate to 7000
[ 9151.900238] perf: interrupt took too long (31833 > 31657), lowering kernel.perf_event_max_sample_rate to 6000
[13024.406999] perf: interrupt took too long (40010 > 39791), lowering kernel.perf_event_max_sample_rate to 4000
[19542.371362] perf: interrupt took too long (50099 > 50012), lowering kernel.perf_event_max_sample_rate to 3000
[30380.717133] perf: interrupt took too long (62978 > 62623), lowering kernel.perf_event_max_sample_rate to 3000
[30380.729842] perf: interrupt took too long (78753 > 78722), lowering kernel.perf_event_max_sample_rate to 2000
[30380.784735] perf: interrupt took too long (98507 > 98441), lowering kernel.perf_event_max_sample_rate to 2000
[30381.244220] perf: interrupt took too long (124173 > 123133), lowering kernel.perf_event_max_sample_rate to 1000
[30384.253023] INFO: NMI handler (perf_event_nmi_handler) took too long to run: 361908.001 msecs
[30385.536514] INFO: NMI handler (perf_event_nmi_handler) took too long to run: 999247.001 msecs
[30385.905688] perf: interrupt took too long (158684 > 155216), lowering kernel.perf_event_max_sample_rate to 1000

2，/proc/cpuinfo

processor    : 0
vendor_id    : GenuineIntel
cpu family    : 6
model        : 45
model name    : Intel(R) Xeon(R) CPU E5-2430 0 @ 2.20GHz
stepping    : 7
microcode    : 0x710
cpu MHz        : 183.906
cache size    : 15360 KB

可以看到cpu MHz  183  呵呵，缩小了十倍，服务器能不慢么!!!!!!