Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I have a RHEL 5.5 on VMWARE running with 64GB Memory. When we untar a file (part of a daily process) the CPU Load sometimes spikes - rendering the server essentially nonresponsive. The disks are thick provisioned so it shouldn't be an issue of needing to dynamically expand the disk.

I have no access to the underlying hardware but need to find out what is causing this since these are database servers that really shouldn't be non-responsive for 30 minutes at a time.

Here is what the load looked like during the untar (note that the cron job apparently got backed up between 20:56 and 21:25) The untar started at 20:51 and ended at 21:02

 top - 20:50:02 up 4 days, 21:26,  0 users,  load average: 0.62, 0.62, 0.75
 top - 20:52:01 up 4 days, 21:28,  0 users,  load average: 4.16, 1.48, 1.02
 top - 20:54:02 up 4 days, 21:30,  0 users,  load average: 11.28, 5.13, 2.41
 top - 20:56:07 up 4 days, 21:33,  0 users,  load average: 14.44, 8.57, 4.02
 top - 21:25:29 up 4 days, 22:02,  0 users,  load average: 154.01, 139.20, 99.18
 top - 21:25:30 up 4 days, 22:02,  0 users,  load average: 154.01, 139.20, 99.18
 top - 21:25:30 up 4 days, 22:02,  0 users,  load average: 154.01, 139.20, 99.18
 top - 21:25:35 up 4 days, 22:02,  0 users,  load average: 159.38, 140.56, 99.84
 top - 21:25:36 up 4 days, 22:02,  0 users,  load average: 163.67, 141.76, 100.44
 top - 21:25:36 up 4 days, 22:02,  0 users,  load average: 163.67, 141.76, 100.44
 top - 21:25:37 up 4 days, 22:02,  0 users,  load average: 163.67, 141.76, 100.44
 top - 21:25:37 up 4 days, 22:02,  0 users,  load average: 163.67, 141.76, 100.44
 top - 21:25:35 up 4 days, 22:02,  0 users,  load average: 159.38, 140.56, 99.84
 top - 21:25:39 up 4 days, 22:02,  0 users,  load average: 163.67, 141.76, 100.44
 top - 21:25:39 up 4 days, 22:02,  0 users,  load average: 163.67, 141.76, 100.44
 top - 21:25:41 up 4 days, 22:02,  0 users,  load average: 171.47, 143.74, 101.31
 top - 21:25:42 up 4 days, 22:02,  0 users,  load average: 171.47, 143.74, 101.31
 top - 21:26:02 up 4 days, 22:02,  0 users,  load average: 137.27, 137.85, 100.28
 top - 21:28:01 up 4 days, 22:04,  0 users,  load average: 21.04, 94.49, 89.05
 top - 21:30:02 up 4 days, 22:06,  1 user,  load average: 3.09, 63.30, 78.28
 top - 21:32:01 up 4 days, 22:08,  1 user,  load average: 2.17, 43.05, 69.04
 top - 21:34:02 up 4 days, 22:10,  1 user,  load average: 0.79, 29.08, 60.77
 top - 21:36:02 up 4 days, 22:12,  1 user,  load average: 0.90, 19.76, 53.51
 top - 21:38:01 up 4 days, 22:14,  1 user,  load average: 0.71, 13.45, 47.10

cat /proc/interrupts shows the following:

           CPU0       CPU1       CPU2       CPU3
  0:  469063418          0          0          0    IO-APIC-edge  timer
  1:        131         64          0          0    IO-APIC-edge  i8042
  6:          5          0          0          0    IO-APIC-edge  floppy
  7:          0          0          0          0    IO-APIC-edge  parport0
  8:          1          0          0          0    IO-APIC-edge  rtc
  9:          0          0          0          0   IO-APIC-level  acpi
 12:        428          8          0        140    IO-APIC-edge  i8042
 15:    3577805     630018       6354        720    IO-APIC-edge  ide1
 51:    3158220     744800      30127      12814   IO-APIC-level  ioc0
 67:   20134707    6847632          0    5355226         PCI-MSI  eth0
 83:          0          0          0          0         PCI-MSI  vmci
NMI:          0          0          0          0
LOC:  469060990  469075818  469075515  469075350
ERR:          0
MIS:          0
share|improve this question
add comment

1 Answer 1

Compression/decompression (like encryption/decryption) is a CPU-bound operation, all else being equal. You're seeing CPU spikes because what you're doing puts a heavy load on the CPU.

If this is making your servers unresponsive, you need to start considering CPU priority or maximum CPU utilization when you run this task and limiting the priority of the untar process.

share|improve this answer
    
The simple act of untar'ing a file should not cause a load to jump to 171+ (and it doesn't do it every time). –  user739866 Apr 2 '13 at 16:46
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.