Please post your server and load averages.

In reply to:

---

[look]$ uptime
09:39:56 up 14 days, 5:40, 9 users, 55.69, 48.82, 64.08

---

I think this is excessive even for shared hosting. In my previous experiences with shared webhosting load averages, in peak times, remained under 10 on a dual processor machine (which still may be excessive).

I don't know how many processors these machines have, but I can determine that the performance of my site is currently far below par with the state of this server. Page serving rates are dropping to and below dial-up speed.

I am interested to see what load averages other customers of DreamHost are experiencing at different times. Please post your results here. :)

Might be beneficial to run these commands to compare everyone's servers:

uptime
cat /proc/cpuinfo
cat /proc/meminfo

Server Name: Rollo
Uptime
09:54:26 up 25 days, 19:57, 4 users, load average: 36.97, 19.01, 12.64
CPU Info
Two Dual Core AMD Opteron(tm) Processors at 2194.612MHz
Memory Info
Total: 4172mb
Used: 4008mb
Free: 164mb



Check out Gordaen's Knowledge, the blog, and the MR2 page.

tank 10:04:14 up 1 day, 23:32, 3 users, load average: 20.71, 12.16, 8.97

(single) Intel(R) Xeon(TM) CPU 2.00GHz
MemTotal: 1031424 kB
MemFree: 36880 kB

Wow, I had never checked before. I have a dual 2.8 Xeon in my garage...

GiraffeDesign.com

[goober]$ uptime
13:16:13 up 14 days, 15:44, 6 users, load average: 6.36, 4.81, 4.45

Two Dual Core AMD Opteron(tm) Processor 175
MemTotal: 4075060 kB
MemFree: 119224 kB

It looks good, but my site has been running very slow since I transferred to DH from ipowerweb few weeks ago.

Leo

server: shemp
cpu: single-processor p3-800 (used to be a dual - dh: where did the other processor go?)
mem: 1.5GB
uptime:
13:59:34 up 22:24, 3 users, load average: 33.92, 22.29, 14.77

servers are seriously overloaded.

apache bench numbers for serving a php-gallery index page (version 1):
Concurrency Level: 1
Time taken for tests: 27.081 seconds
Complete requests: 1
Failed requests: 0
Broken pipe errors: 0
Total transferred: 69832 bytes
HTML transferred: 69351 bytes
Requests per second: 0.04 [#/sec] (mean)
Time per request: 27081.00 [ms] (mean)
Time per request: 27081.00 [ms] (mean, across all concurrent requests)
Transfer rate: 2.58 [Kbytes/sec] received

the same with wget:
HTTP request sent, awaiting response... 200 OK
Length: unspecified [text/html]

[ <=> ] 69,351 3.80K/s

13:58:30 (3.80 KB/s) - `index.php' saved [69351]

real 36.455 user 0.030 sys 0.010 pcpu 0.10

Server Name: dontddosme

Uptime
15:01:26 up 26 days, 1:06, 5 users, load average: 3.99, 2.21, 1.72 Yeah, baby!

CPU Info
Two Dual Core AMD Opteron(tm) Processors at 2194 MHz

Memory Info

Total: 4173 MB
Used: 3994 MB
Free: 179 MB

You are welcome to a refund.-Michael;Comparing 'DH of today' with 'DH of old' does not serve much purpose.-Dallas

[galaga]$ uptime
07:23:51 up 1 day, 10:54, 3 users, load average: 4.90, 3.75, 3.77

cpuinfo lists 4 Intel(R) Xeon(TM) CPU 2.40GHz

Mem: total: 1586663424
used: 1531564032
free: 55099392




--Matttail
art.googlies.net - personal website

my sites are dead:

[twinkie]$ uptime
09:49:53 up 8 days, 21:36, 5 users, load average: 55.05, 33.22, 28.92

[twinkie]$ cat /proc/cpuinfo
processor : 0
vendor_id : GenuineIntel
cpu family : 15
model : 2
model name : Intel(R) Xeon(TM) CPU 2.40GHz
stepping : 9
cpu MHz : 2400.113

[twinkie]$ cat /proc/meminfo
total: used: free: shared: buffers: cached:
Mem: 2117214208 2045505536 71708672 0 6070272 730046464
Swap: 6465036288 913444864 5551591424
MemTotal: 2067592 kB
MemFree: 70028 kB
MemShared: 0 kB
Buffers: 5928 kB
Cached: 654180 kB
SwapCached: 58756 kB
Active: 205964 kB
Inactive: 513360 kB
HighTotal: 1179584 kB
HighFree: 2732 kB
LowTotal: 888008 kB
LowFree: 67296 kB
SwapTotal: 6313512 kB
SwapFree: 5421476 kB

[twinkie]$ uptime
09:54:28 up 8 days, 21:41, 5 users, load average: 25.97, 31.66, 29.67

finally got into SSH after about 20 minutes.

[crunch]$ uptime
09:58:57 up 1 day, 21:13, 3 users, load average: 49.73, 57.90, 62.96

uhoh. Apparently it's too hard to find out what the cpuinfo is... hanging on that request.

And a prompt. No response, just a prompt a couple of minutes later.

Ran the request again, came up right away :

[crunch]$ cat /proc/cpuinfo
processor : 0
vendor_id : AuthenticAMD
cpu family : 15
model : 35
model name : Dual Core AMD Opteron(tm) Processor 175
stepping : 2
cpu MHz : 2194.552
cache size : 1024 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov
pat pse36 clflush mmx fxsr sse sse2 ht syscall mmxext lm 3dnowext 3dnow pni
bogomips : 4377.80

processor : 1
vendor_id : AuthenticAMD
cpu family : 15
model : 35
model name : Dual Core AMD Opteron(tm) Processor 175
stepping : 2
cpu MHz : 2194.552
cache size : 1024 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov
pat pse36 clflush mmx fxsr sse sse2 ht syscall mmxext lm 3dnowext 3dnow pni
bogomips : 4377.80


[crunch]$ cat /proc/meminfo
total: used: free: shared: buffers: cached:
Mem: 4172877824 4044550144 128327680 0 12095488 1840893952
Swap: 6465036288 65224704 6399811584
MemTotal: 4075076 kB
MemFree: 125320 kB
MemShared: 0 kB
Buffers: 11812 kB
Cached: 1796372 kB
SwapCached: 1376 kB
Active: 506380 kB
Inactive: 1303740 kB
HighTotal: 3211200 kB
HighFree: 3792 kB
LowTotal: 863876 kB
LowFree: 121528 kB
SwapTotal: 6313512 kB
SwapFree: 6249816 kB


[crunch]$ uptime
10:06:35 up 1 day, 21:21, 2 users, load average: 48.52, 48.99, 56.18


*edit*
Not sure why my info is coming through so much more detailed than everyone elses.

here's the top bit of Top:

[crunch]$ top
top - 10:10:20 up 1 day, 21:25, 2 users, load average: 55.85, 50.45, 55.01
Tasks: 4 total, 2 running, 2 sleeping, 0 stopped, 0 zombie
Cpu(s): 81.7% user, 18.3% system, 0.0% nice, 0.0% idle
Mem: 4075076k total, 3924080k used, 150996k free, 10832k buffers
Swap: 6313512k total, 63696k used, 6249816k free, 1720844k cachedEdited by jcovergaar on 08/24/06 10:11 AM (server time).

11:49:15 up 1:40, 4 users, load average: 19.18, 28.38, 29.66

CPU: 3 Intel(R) Xeon(TM) CPU 2.80GHz
mem:

[look]$ cat /proc/meminfo
total: used: free: shared: buffers: cached:
Mem: 3178242048 1608818688 1569423360 0 21331968 458686464
Swap: 6465036288 0 6465036288
MemTotal: 3103752 kB
MemFree: 1532640 kB
MemShared: 0 kB
Buffers: 20832 kB
Cached: 447936 kB
SwapCached: 0 kB
Active: 180424 kB
Inactive: 288780 kB
HighTotal: 2228096 kB
HighFree: 783240 kB
LowTotal: 875656 kB
LowFree: 749400 kB
SwapTotal: 6313512 kB
SwapFree: 6313512 kB

[soto]$ uptime
12:02:11 up 17:58, 3 users, load average: 91.93, 70.12, 49.84

cpuinfo lists 4 Intel(R) Xeon(TM) CPU 2.80GHzz

Mem: total: 3177984000
used: 2955640832
free: 222343168

Melrose:
13:00:06 up 2:42, 8 users, load average: 22.19, 32.91, 37.51

Sites taking ~5 minutes to load

It should be noted that all of these uptime reports of not accurate. Becuase of the ongoing network issues loads are higher than normal on many servers becuase they are not able to properly connect as they're expecting to. This drives up the load rating on the machine, and even makes it inacurate becuase the server is waiting on a response from the network (which is delayed) though it's not really processing anything while it's waiting for the response.

At the worst of the network issues before, loads got upwards of 200 on some servers. A server would normally crash if load was truly that high, but they were still chugging along - waiting on the network.

I would strongly suggest that you re-evaluate server loads once DH has their network issues resolved.



--Matttail
art.googlies.net - personal website

once DH has their network issues resolved.

How so? The "network issues" have plagued DH for many months. So testing them during that period is completely accurate. This is because the tests are being run during a period of time in reality, unlike the hypothetical "when DH has their network issues resolved" time period you propose.

Ok, so that's mostly in jest, though it's true.


GiraffeDesign.com

mmm ... not sure how long this will last ... but happy - happy - joy - joy!!

[twinkie]$ uptime
14:30:23 up 9 days, 2:16, 3 users, load average: 2.40, 2.63, 3.30

can't describe relief...

I almost hate to post this, given the numbers others have posted:

[genki]$ uptime
14:48:47 up 17 days, 8:27, 4 users, load average: 0.28, 0.48, 0.56

In reply to:

---

How so? The "network issues" have plagued DH for many months. So testing them during that period is completely accurate. This is because the tests are being run during a period of time in reality, unlike the hypothetical "when DH has their network issues resolved" time period you propose.

Ok, so that's mostly in jest, though it's true.

---

Yeah, you're right. I just wanted to make sure people realized that the numbers were not necessarly a *norm*. It's also worth pointing out that the numbers are actually a little off becuase of the 'network issues'. Processes are queueing up waiting for a response from the network, so the numbers grow even though the CPU isn't actually doing anything.



--Matttail
art.googlies.net - personal website

In reply to:

---

Yeah, you're right. I just wanted to make sure people realized that the numbers were not necessarly a *norm*. It's also worth pointing out that the numbers are actually a little off becuase of the 'network issues'. Processes are queueing up waiting for a response from the network, so the numbers grow even though the CPU isn't actually doing anything.

---

Yeah, especially when the numbers I just got are outrageous.

tank:~$ uptime
22:21:51 up 3 days, 11:49, 4 users, load average: 810.16, 480.42, 215.55



GiraffeDesign.com

In reply to:

---

tank:~$ uptime
22:21:51 up 3 days, 11:49, 4 users, load average: 810.16, 480.42, 215.55

---

Good Grief! Is that Scotty I hear screaming in the background, "I'm givin' 'er all she's got, Cap'n!" that is messed up. :-(

--rlparker

It's gotta be a "tank" to run through that!

Check out Gordaen's Knowledge, the blog, and the MR2 page.

personal account:
zangief:~$ uptime
09:14:43 up 17 days, 21:11, 2 users, load average: 1.22, 1.81, 2.00

business account:
[gaz]$ uptime
09:16:43 up 27 days, 18:31, 2 users, load average: 1.46, 1.87, 2.10

[hillhurst]
Uptime 10:26:39 up 27 days, 20:30, 2 users, load average: 1.12, 1.32, 1.49
CPU Info Two Dual Core AMD Opteron Processors at 2194.592
Memory Info
Total: 4173 mb
Used: 4038 mb
Free: 134 mb

[twinkie]
Uptime 10:40:30 up 9 days, 22:27, 6 users, load average: 3.32, 3.81, 4.69
CPU Info Four Intel(R) Xeon(TM) CPU 2.40GHz at 2400.113
Memory Info
Total: 2117 mb
Used: 2056 mb
Free: 609 mb

Things seem to be looking up now; I've just been moved from Wilmington to Cichlet, which seems to have more users, but lesser resource utilisation:

17:16:46 up 27 days, 20:29, 9 users, load average: 1.43, 1.42, 1.62



======================
Visit my blog
(http://pkchukiss.sgblogging.com)

[homer]$ uptime
18:48:40 up 16 min, 2 users, load average: 0.03, 0.03, 0.01
cpu info:
3 x Intel(R) Xeon(TM) CPU 2.80GHz
MemTotal: 2067388 kB
MemFree: 1866740 kB

[look]$ uptime
08:32:59 up 10 min, 2 users, load average: 1.16, 1.31, 0.72

That didn't last.

[look]$ uptime
08:54:49 up 11 min, 3 users, load average: 56.46, 20.33, 7.63

I guess not! Must have been a "rough" 22 minutes, eh!
--rlparker

So I checked my site today and it was amazingly quick. For about 3 minutes it was as if I was the only person using the network. I logged into the shell and found that the server I was on had just rebooted. I monitored activity over the next 20 minutes...

[look]$ uptime
11:32:28 up 1 min, 1 user, load average: 0.82, 0.22, 0.07
[look]$ uptime
11:33:39 up 2 min, 1 user, load average: 1.97, 0.67, 0.24
[look]$ uptime
11:36:07 up 5 min, 1 user, load average: 3.55, 1.74, 0.69
[look]$ uptime
11:43:14 up 12 min, 1 user, load average: 8.42, 3.39, 1.65
[look]$ uptime
11:43:19 up 12 min, 1 user, load average: 8.31, 3.45, 1.67
[look]$ uptime
11:43:21 up 12 min, 1 user, load average: 8.31, 3.45, 1.67
[look]$ uptime
11:43:31 up 12 min, 1 user, load average: 9.42, 3.86, 1.83
[look]$ uptime
11:43:32 up 12 min, 1 user, load average: 10.91, 4.26, 1.97
[look]$ uptime
11:43:36 up 12 min, 1 user, load average: 10.91, 4.26, 1.97
[look]$ uptime
11:43:37 up 12 min, 1 user, load average: 13.16, 4.84, 2.17
[look]$ uptime
11:44:39 up 13 min, 1 user, load average: 37.16, 12.42, 4.89
[look]$ uptime
11:46:34 up 15 min, 1 user, load average: 113.45, 46.91, 18.24
[look]$ uptime
11:47:00 up 15 min, 1 user, load average: 134.36, 60.61, 24.06
[look]$ uptime
11:48:39 up 17 min, 1 user, load average: 162.60, 91.96, 39.23
[look]$ uptime
11:51:51 up 20 min, 1 user, load average: 184.55, 133.92, 64.71

As I watched these load averages climb I also observed a proportional rise in the delay of server response. Ranging between 5 to even more than 60 seconds to respond (even the shell demonstrates these extremes in latency). This is obviously /completely/ useless for serving pages to my users. But what I want to know is, what is causing this and who is affected? I can access other dreamhost sites very quickly (and this is consistent), but mine suffers. Is it just the box I'm on? Is it a section of the network I'm on? Should I request to be moved?

Are dreamhost users abusing their server privileges or is this a network issue? I've only been with dreamhost for 4 days now, perhaps my timing was poor but I have yet to enjoy any decent service. I hope this is resolved so I can feel good about the money I spent and go on to recommend dreamhost to my friends.

In reply to:

---

This is obviously /completely/ useless for serving pages to my users. But what I want to know is, what is causing this and who is affected?

---

If your loads are that high it's clearly NFS. What the problem is? Only DH would know for sure but it's the network or the load on the NFS is too high.

digitalrundown.com

I think the (probably accurrate) assumption here is that those posting reports to this thread know what the numbers they are seeing mean. That said, I'm *not* so sure that is true for many who may be reading this thread, so, in the interest of "general education", here a couple of links for any who may be curious about how to make sense of these numbers:

General discussion of "Overloading"

"Server Loads Explained"

Of course, these articles represent the opinions and beliefs of their authors, and as to the relevance of their belilefs to your server's performance, YMMV. All the standard disclaimers apply. There a "lots" of these types of articles available on the web; I just thought these were representative and they might be of interest.

--rlparker

That's terrible!!!!! No server uptime exceed 30 days!

Here are my Hosting Plan Comparision and $97 off DreamHost Promo Code

How do you go about finding these statistics?

I think I've got some of you beat, check these out:

[starburst]$ w
19:28:33 up 2:37, 2 users, load average: 854.79, 922.14, 762.35
USER TTY FROM LOGIN@ IDLE JCPU PCPU WHAT
[starburst]$ uptime
19:29:11 up 2:37, 2 users, load average: 506.41, 824.97, 736.45
[starburst]$ uptime
19:29:14 up 2:37, 2 users, load average: 475.23, 813.21, 733.12
[starburst]$ uptime
19:29:16 up 2:37, 2 users, load average: 475.23, 813.21, 733.12
[starburst]$ uptime
19:29:17 up 2:37, 2 users, load average: 475.23, 813.21, 733.12
[starburst]$ uptime
19:30:08 up 2:38, 2 users, load average: 251.78, 701.15, 699.25
[starburst]$ uptime
19:30:09 up 2:38, 2 users, load average: 232.58, 689.71, 695.56
[starburst]$ uptime
19:31:04 up 2:39, 2 users, load average: 94.58, 574.21, 655.73
[starburst]$ uptime
19:31:05 up 2:39, 2 users, load average: 94.58, 574.21, 655.73
[starburst]$ uptime
19:31:31 up 2:40, 2 users, load average: 64.17, 528.51, 638.44
[starburst]$ uptime
19:34:45 up 2:43, 2 users, load average: 84.60, 307.57, 529.64
19:34:45 up 2:43, 2 users, load average: 84.60, 307.57, 529.64
[starburst]$ w
19:35:26 up 2:43, 2 users, load average: 100.04, 283.52, 512.18
USER TTY FROM LOGIN@ IDLE JCPU PCPU WHAT
[starburst]$ uptime
19:35:33 up 2:44, 2 users, load average: 102.20, 280.93, 510.11
[starburst]$ uptime
19:35:34 up 2:44, 2 users, load average: 104.27, 278.39, 508.06
[starburst]$ uptime
19:35:35 up 2:44, 2 users, load average: 104.27, 278.39, 508.06
[starburst]$ uptime
19:35:35 up 2:44, 2 users, load average: 104.27, 278.39, 508.06

.. a few minutes later..
19:40:31 up 2:49, 2 users, load average: 204.26, 222.52, 419.98

.. and now my shell is frozen.

and guess what? This is not uncommon. Starburst averages 200-400 regularly. I emailed support asking about moving to a different server:

>Would upgrading to the next level of service get us on a better server?

No. Upgrading plans doesn't move you to a different server. Let us
know if you have any further questions.
Thanks!
Dan

I was told new hardware was coming "next week". Strange. 400 server load doesn't just happen one day.