Jump to content

Server Discussion


SDS

Recommended Posts

  • Replies 55
  • Created
  • Last Reply

Top Posters In This Topic

I/O read and Writes

Memory utilization

Physical Disk\Avg. Disk Reads/sec

Physical Disk\Avg. Disk Writes/sec

machine processor time

SQL processor time

 

 

Would be a good place to start, Monitor every 5 minutes for a day and see where things settle out on first pass. don't add too much monitoring at once you will not get any numbers that are useful.

Also location probably will not play much into performance unless one site has a much larger bandwidth. The Hard drives personally I would stick with 2 drive's in a raid one config, you can still backup but if one drive fails you're not down and the raid will rebuild the other drive once replaced, Backups are better for corruption than Hardware issues. Make sure your always installing service packs and security patches (Microsoft every 2nd Tuesday of the month)

Oh the other part you can NEVER have too much memory, but the OS has to be able to support it, what OS is the server running/ what are you going to use on the new server?

Edited by gordong
Link to comment
Share on other sites

07:40:01 PM CPU %user %nice %system %iowait %steal %idle

07:50:01 PM all 32.38 0.00 14.83 1.42 0.00 51.38

08:00:01 PM all 29.81 0.06 14.97 0.97 0.00 54.18

08:10:01 PM all 26.41 0.00 12.45 0.76 0.00 60.39

08:20:01 PM all 25.43 0.00 11.53 0.95 0.00 62.09

08:30:01 PM all 25.44 0.00 10.80 1.15 0.00 62.62

08:40:01 PM all 26.13 0.00 10.55 0.51 0.00 62.81

08:50:01 PM all 23.92 0.00 7.62 0.54 0.00 67.92

09:00:01 PM all 22.59 0.06 8.42 0.53 0.00 68.39

09:10:01 PM all 22.00 0.00 8.05 0.90 0.00 69.04

09:20:01 PM all 23.27 0.00 8.06 0.63 0.00 68.04

09:30:01 PM all 19.52 0.00 6.82 0.91 0.00 72.75

09:40:01 PM all 20.83 0.00 7.00 0.40 0.00 71.77

09:50:01 PM all 17.82 0.00 6.41 0.38 0.00 75.38

10:00:01 PM all 18.58 0.06 7.11 0.38 0.00 73.86

10:10:01 PM all 17.74 0.00 6.40 0.43 0.00 75.43

Average: all 23.46 0.01 9.40 0.72 0.00 66.40

 

load at 7:40 was around 15 or so. Several seconds to execute each page.

 

top - 21:19:07 up 204 days, 19:27, 1 user, load average: 3.97, 3.18, 3.02

Tasks: 262 total, 6 running, 250 sleeping, 0 stopped, 6 zombie

Cpu(s): 27.9%us, 18.5%sy, 0.0%ni, 52.2%id, 1.2%wa, 0.0%hi, 0.1%si, 0.0%st

Mem: 12302712k total, 6826720k used, 5475992k free, 164572k buffers

Swap: 2096472k total, 476868k used, 1619604k free, 3324168k cached

 

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND

3583 mysql 15 0 2653m 587m 3980 S 269.0 4.9 102021:54 mysqld

8684 apache 15 0 588m 59m 15m S 10.6 0.5 0:00.64 httpd

8658 apache 15 0 592m 63m 18m S 10.3 0.5 0:01.12 httpd

8678 apache 15 0 558m 38m 22m S 7.0 0.3 0:00.84 httpd

8566 apache 15 0 590m 78m 30m R 6.0 0.7 0:05.20 httpd

8353 apache 16 0 594m 105m 55m S 3.7 0.9 0:11.71 httpd

Link to comment
Share on other sites

there is something wrong with your system

Based on what? Care to elaborate?

 

On a side note, this may have been brought up elsewhere but have you considered spinning up a box in Amazon Web Services, Google cloud or some other? Too much coin? Pretty easy to scale when demand warrants it.

Link to comment
Share on other sites

Based on what? Care to elaborate?

Although I'm not a sysadmin, I'm in the web development business since 1999 as lead developer, and I can safely say there is a problem when your load average is more than 0.20 in normal conditions and swap used shows too something wrong.

 

so, and assuming SDS posted normal conditions and not peak conditions :

 

 

load average: 3.97, 3.18, 3.02

means for me that your CPU is too heavily used

 

 

Tasks: 262 total, 6 running, 250 sleeping, 0 stopped, 6 zombie

you should have 0 zombie process ; some proccess might have been wrongly terminated

 

 

Cpu(s): 27.9%us, 18.5%sy, 0.0%ni, 52.2%id, 1.2%wa, 0.0%hi, 0.1%si, 0.0%st

way too much on normal conditions

 

 

Mem: 12302712k total, 6826720k used, 5475992k free, 164572k buffers

for me, it's fine

 

 

Swap: 2096472k total, 476868k used, 1619604k free, 3324168k cached

means swap has been used (maybe once, aybe more), and generally it's a bad thing, server didn't have enough RAM, so went into swap (physical disk allocated memory), and then everything runs more slowly and everything become a vicious circle.

 

 

3583 mysql 15 0 2653m 587m 3980 S 269.0 4.9 102021:54 mysqld

under normal conditions (no cron used by the time, no heavy requests on mysql by the time), that's way too much.

 

that said, SDS said he changed something on mysql (internal stats disabled ?), it could have done the trick.

Edited by Repulsif
Link to comment
Share on other sites

Ok, this is what I'm looking at:

 

E3-1270v2 (for a one-time fee I can upgrade to version V3 of this)
32GB DDR-3 ECC RAM
RAID controller: LSI 9271-4i
4X 240GB SSD (m500 or intel520, RAID 10 for main OS and DB)
1x 120GB SSD in ssd slot, directly connected to motherboard for backups
100mbit unmetered or 30TB on 1gig port
13x IPv4
IPMI/KVM, firewalled

 

​This thing would fly I think.

Link to comment
Share on other sites

×
×
  • Create New...