Web Hosting Talk : Web Hosting Main Forums : Hosting Security and Technology : May switch cause high load on powerful server?

[Driimer]

Aight, I have two quad core processors and load is like 15.
May it be caused by switch if it doesnt let traffic trough properly?

if dmesg grep eth shows 100 full duplex is it normal or should it be 1000 full duplex? how can I make it 1000 full duplex on centos 5 ?

0000:0a:02.0: eth0: (PCI Express:2.5GB/s:Width x4)
0000:0a:02.0: eth0: Intel(R) PRO/1000 Network Connection
0000:0a:02.0: eth0: MAC: 3, PHY: 5, PBA No: ffffff-0ff
0000:0a:02.0: eth1: (PCI Express:2.5GB/s:Width x4)
0000:0a:02.0: eth1: Intel(R) PRO/1000 Network Connection
0000:0a:02.0: eth1: MAC: 3, PHY: 5, PBA No: ffffff-0ff
ADDRCONF(NETDEV_UP): eth0: link is not ready
0000:0a:02.0: eth0: Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX
0000:0a:02.0: eth0: 10/100 speed: disabling TSO
ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
ADDRCONF(NETDEV_UP): eth1: link is not ready
0000:0a:02.0: eth1: Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX
0000:0a:02.0: eth1: 10/100 speed: disabling TSO
ADDRCONF(NETDEV_CHANGE): eth1: link becomes ready
eth0: no IPv6 routers present
eth1: no IPv6 routers present
ADDRCONF(NETDEV_UP): eth0: link is not ready
0000:0a:02.0: eth0: Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX
0000:0a:02.0: eth0: 10/100 speed: disabling TSO
ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
ADDRCONF(NETDEV_UP): eth1: link is not ready
0000:0a:02.0: eth1: Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX
0000:0a:02.0: eth1: 10/100 speed: disabling TSO
ADDRCONF(NETDEV_CHANGE): eth1: link becomes ready
eth0: no IPv6 routers present
eth1: no IPv6 routers present

[beastserv]

Can you provide us with the output of the command " top "

and about your question - actualy a 1000mbps line would make your hardware work harder, due to larger amount of I/O that a 1gbps link can eat up in contrast to a 100mbps link.

[Driimer]

Here you go:

__PID_____USER_____PR__NI__VIRT__RES__SHR_S_%CPU__%MEM____TIME+____COMMAND
_6212_srcds014_____15___0__213m_201m_7404_S___12__1.3_829:25.22 srcds_i686
_4397_srcds017_____15___0__162m_153m_7304_S___11__1.0__59:37.82 srcds_i686
_9766_srcds016_____15___0__130m_120m_7328_S____6__0.7__43:49.45 srcds_i686
_8646_srcds010_____15___0__111m__98m_7264_S____3__0.6__15:33.58 srcds_i686
_4692_srcds006_____15___0__189m_178m_7428_S____2__1.1__54:15.96 srcds_i686
15711_srcds011_____15___0__116m_103m_7456_S____2__0.6__20:49.17 srcds_i686
15934_srcds005_____15___0__103m__90m_7284_S____2__0.6__12:38.15 srcds_i686
_8318_srcds007_____15___0__115m_101m_7336_S____1__0.6__70:52.09 srcds_i686
17876_root_________15___0_12740_1172__804_R____1__0.0___0:00.12 top
_2157_srcds003_____15___0_96752__79m_6916_S____0__0.5___9:37.12 srcds_i686
_6678_srcds012_____15___0__119m_106m_7472_S____0__0.7__17:31.71 srcds_i686
_7016_srcds002_____15___0__148m__78m__12m_S____0__0.5___2:10.72 srcds_i686
_8099_srcds008_____15___0_78112__66m_7372_S____0__0.4__28:09.32 srcds_i686
11173_srcds015_____15___0__101m__88m_7420_S____0__0.6__10:49.15 srcds_i686
12067_srcds004_____15___0__127m_113m_7432_S____0__0.7__20:20.02 srcds_i686
24526_srcds013_____15___0_97932__81m_7160_R____0__0.5___9:25.22 srcds_i686
25848_srcds009_____15___0_73388__61m_7192_S____0__0.4___2:01.61 srcds_i686
27002_srcds001_____15___0_98592__81m_6916_S____0__0.5__10:37.66 srcds_i686

So it can't be switch and patch cat5e utp cables fault?

[beastserv]

nope its not your cable, you got plenty of counter strike servers running there.
how many slots per server?

a load of 15 sound Normal for such kind of server usage.

in this page alone we can see about 17 game servers.

lets say 12players per server ?

17X12 ?
you are overkilling your hardware mate.

how much ram do you have there?

show us the output of " free -m "
and " iostat "
which kind of disks are you useing?
have you disabled un wanted services / kernel modules ?

[Driimer]

16gb ram

____________________total_______used______free____shared_____buffers___cached
Mem:________________16051______15662_______389_______0_________528_______10810
-/+ buffers/cache:_____4323_______11727
Swap:_______________1983___________0_____1983

Adding ram doesnt seem to help since same result was with 8gb.

avg-cpu:__%user___%nice_%system_%iowait__%steal___%idle
___________2.15___0.03____0.82____0.56____0.00___96.44

Device:________tps______Blk_read/s___Blk_wrtn/s___Blk_read___Blk_wrtn
sda__________16.95________16.24_______410.86___26647872__673999262
dm-0_________51.84________16.24_______410.86___26643882__673998912
dm-1__________0.00_________0.00_________0.00________896________328

15krpm discs

Haven't disabled any kernel modules nor services.

Would changing from raid 1 into raid 0 system make huge difference in server load?

[beastserv]

Originally Posted by Driimer

16gb ram



Adding ram doesnt seem to help since same result was with 8gb.



15krpm discs

haven't disabled any kernel modules nor services.

you should fine tune your machine, get rid of un-used modules and services.

another thing, try looking at /var/log/messages for any kind of errors.

...split your game servers to 2 machines. (ouch i know).

game servers eat a lot, and im pretty sure they are loaded with players.

[Driimer]

Would changing from raid 1 into raid 0 system make huge difference in server load?

Repeating it since you replied before I edited. I know I have to format discs if I make change from raid 0 into raid 1 but does it format discs if I make it from raid 1 into raid 0?

Edit: that cat5e is my cable there tho, same goes to switch and that's why I think they aren't proper tech over there.

[beastserv]

RAID 0 is not a good option for production servers, as if one disk get damaged you loose your whole disk array.

stick with raid1 , for 15K disks its a minor diffrance in contrast to the benefit you gain while mirroring.

anyway, first of all , get rid of services you dont need and tweak your kernel a bit.
that could help just a bit more, but as it seems, you are asking your hardware for too much...