Stewart Stremler wrote:
> begin quoting Andrew Lentvorski as of Thu, Feb 15, 2007 at 06:07:41PM
> -0800:
>> Neil Schneider wrote:
>>
>> >I wish more people would participate in the discussion. I don't
>> think
>> >we need to rush, but sometime in the next month would probably be a
>> >good time for coming to a consensus.
>>
>> I agree. A little more participation would be nice.
>
> Heh.
>
> Have we run any profiling tools on the existing server to determine
> where the real bottlenecks are, rather than speculating?
We're so far into swap, that it's hard to tell where other problems
could be.
sparkplug:~# cat /proc/meminfo
total: used: free: shared: buffers: cached:
Mem: 512229376 504270848 7958528 0 14385152 223092736
Swap: 1768902656 297107456 1471795200
I don't know how much more profiling anyone else has done. I'm not
even sure the tools are in place, to do the job.
> In general, oodles of ECC RAM and good I/O between it and the
> processor are where I'd tend to prefer to emphasis be placed.
> I haven't found any discussion about if SATA is as good as
> low-processor-impact as SCSI, although Tracy and I discussed
> CTQ and NTQ (?) on the channel the other day for a bit.
>From what I've read, the command queing that gave scsi an advantage is
now available in SATA-II drives, making them a better price
performance value.
> Indeed.
>
> A new server would be okay. What are our funds like?
According to our treasurer
http://www.kernel-panic.org/Members/ggeller/ggeller we have $5795.14.
So $2,000 would still leave us with enough to have a super picnic this
year. :-)
> Going uber-cheap-but-better-than-what-we-have-now is nice, and
> spending decent money for a really good machine is nice; I'm not
> so much happy with a middle-of-the-road approach, however. Unlike
> James, I might be happier with $2k spent on a server than $1k.
Our web site always seems to run high resource applications like
plone, so we need lots of horsepower to make it work. We can get a
pretty powerful server in the $2K range. I think that should be the
neighborhood. We need at least 4Gb of memory, since the next iteration
is more likely to use more memory than less, if history is a guide.
> The current server is slow enough to where I don't consider it
> useful[1], so I haven't hit the KPLUG website in ages. So long
> as the replacement is *better* (in performance and at least as
> good in stability) I will make no complaints.
>
> [1] No idea if this is a processor, bus, disk, or software problem.
Well here's more than you probably ever wanted to know about the server.
sar -b # Report I/O and transfer rate statistics
12:01:49 AM tps rtps wtps bread/s bwrtn/s
Average: 19.08 5.37 13.71 180.80 240.40
sparkplug:~# sar -B # Report paging statistics
12:00:21 AM pgpgin/s pgpgout/s fault/s majflt/s
Average: 90.40 120.20 0.00 0.00
sparkplug:~# sar -c # Report process creation activity
12:00:21 AM proc/s
Average: 0.45
sparkplug:~# sar -d # Report activity for
each block device
12:00:21 AM DEV tps rd_sec/s wr_sec/s
Average: dev8-0 19.08 180.80 240.40
sparkplug:~# sar -n DEV # Report network statistics
12:00:21 AM IFACE rxpck/s txpck/s rxbyt/s txbyt/s
rxcmp/s txcmp/s rxmcst/s
Average: lo 3.07 3.07 5223.54 5223.54
0.00 0.00 0.00
Average: eth0 14.80 11.59 2398.49 6468.32
0.00 0.00 0.00
sparkplug:~# sar -P ALL # Report per-processor
statistics
12:00:21 AM CPU %user %nice %system %iowait %idle
Average: all 27.31 0.00 1.64 0.00 71.05
sparkplug:~# sar -q #Report queue length and load averages
12:00:21 AM runq-sz plist-sz ldavg-1 ldavg-5 ldavg-15
Average: 0 83 0.42 0.43 0.43
sparkplug:~# sar -r #Report memory and swap space
utilization statistics
12:00:21 AM kbmemfree kbmemused %memused kbbuffers kbcached
kbswpfree kbswpused %swpused kbswpcad
Average: 5199 495025 98.96 16968 50695
1437328 290116 16.79 158432
sparkplug:~# sar -R # Report memory statistics.
12:00:21 AM frmpg/s bufpg/s campg/s
Average: -0.83 4.34 0.80
sparkplug:~# sar -u #Report CPU utilization
12:00:21 AM CPU %user %nice %system %iowait %idle
Average: all 27.31 0.00 1.64 0.00 71.05
sparkplug:~# sar -v # Report status of inode, file and
other kernel tables
12:00:21 AM dentunusd file-sz inode-sz super-sz %super-sz
dquot-sz %dquot-sz rtsig-sz %rtsig-sz
Average: 1652 1169 1098 0 0.00
0 0.00 0 0.00
sparkplug:~# sar -w # Report system switching activity.
12:00:21 AM cswch/s
Average: 117.23
sparkplug:~# sar -W #Report swapping statistics
12:00:21 AM pswpin/s pswpout/s
Average: 0.95 0.44
--
Neil Schneider pacneil_at_linuxgeek_dot_net
http://www.paccomp.com
Key fingerprint = 67F0 E493 FCC0 0A8C 769B 8209 32D7 1DB1 8460 C47D
A banker is a fellow who lends you his umbrella when the sun is
shining, but wants it back the minute it begins to rain. - Mark Twain
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-steer
