• Please review our updated Terms and Rules here

Forum Issues, Let's work together to determine some details

billdeg

Technician
Joined
Nov 18, 2003
Messages
3,885
Location
Landenberg, PA USA
Hi all,
I have been reading about forum issues and following the threads related. Thanks everyone for reporting in. Problem is, when I use the forum I don't have any slowness. So, I'd like to ask that we use this thread to help me track down the cause of the slowness. Please report issues as follows

1. Date and time of the issue

2. Function effected (for example: My reply post hangs after I submit, it took 45 seconds before the screen refreshed and my post appeared in the list")

3. Basics of your set up (for example: Windows 10 with Firefox 59.0.2)

4. Your location / name of your ISP or upstream provider

NOTE: If you have a lot of tech details beyond this simple 4 point test, for now just hold onto that level of stuff. I am looking for a pattern more so than the details of a ping or tracert.

Thanks
Bill
 
Loading this thread right now took 27 seconds. Actually I don't mean loading and displaying the whole page. It seems like the latency of the initial server response. But I don't have the means to narrow it that well at the moment.

The trouble is though, that my ISP is FreedomPop, over the Sprint LTE network inside an electrically noisy metal building. I can consistently load thread pages on a completely different forum site running the same version of VBulletin faster. However, I don't think it's safe to say that I'm not experiencing latency due to my connection.

Unfortunately I spend most of my time here in this situation. Now I can reasonably guess that the problem is on the server end due to the relative response times of other sites, but that still doesn't guarantee that I'm not having local latency at just the wrong times every time.

Edit: hitting the Post button caused a 34 second delay and a "Stay on this page/leave" prompt. Hitting Stay caused my browser "NB Browser" to go unresponsive to this page; something I've never encountered before here, at all. Despite hitting Stay and then closing the browser, my post was actually posted.
 
FWIW, the forum works fine for me as well, on a variety of operating systems, browsers and devices (real computers with keyboards, none of that tablet or mobile crap!).
 
Please try if you have issues to answer my four questions so I can make a case to the ISP, otherwise we're not going to get what we want. Please refer to the original post, thanks. The running theory at the moment is that the server itself is underpowered. It is also possible that the location of the server (Australia) can be causing lag issues with certain locations that are far away from it. I need to differentiate these two with consistent data, you know computer stuff.
b
 
Bill, a couple of years ago I had a similar performance issue with the website at my $dayjob, and we used Solarwinds Pingdom to help troubleshoot. The pingdom reports include a vast amount of information that can show you the performance, and their tools do a pretty good job of estimating CPU load on the server. I think they have a free 30-day trial, and I found the results in the first 30 days helped me fix the caching/disk issues I was having. I also ran several stress tests with ApacheBench (ab) to help troubleshoot the issue. Due to the network speed I have available to me, I won't run an ab stress test without permission on any server, as that could easily be construed as a DoS attack; but if I had permission I would be glad to help you stress the server (see my speedtest results below for my 'qualifications.')

I have noted some slowness on two different connections with a couple of different devices, but it's intermittent on both. So I can't at the moment answer the time piece of your four questions, but here are the other three for my two connections:

2.) All server interaction slows, and on my CentOS 7 laptop with Firefox I get the 'loading' rotating circle in the tab for a long time (20-45 seconds), and on my Samsung Galaxy S5 Chrome actually hangs and it slows the whole phone down for 10-30 seconds.

3.) Setup: two different clients: Samsung Galaxy S5 phone with Chrome; Dell Precision M6700 Core i7-3740QM with 16GB RAM, running CentOS 7's Firefox 52 ESR default packages, with NoScript loaded.

4.) ISP: three different scenarios, two ISP's:
a: MCNC, 1Gb/s symmetric peak, 100Mb/s committed, Cisco 7609 edge router with dual Supervisor 720's, 1Gb/s ethernet to the desk, hardwired to the laptop. Speedtest.net goes crazy from my laptop, with download speeds typically in the 250Mb/s range and upload speeds in the 750Mb/s range (typical result: http://www.speedtest.net/result/7263354611 ). So it is not a network issue on my end. :)
b: Comporium Communications: from my desk, 15Mb/s bidirectional with a Cisco 7609 edge router. From home, ADSL with 7Mb/s down and .5Mb/s up. Oddly enough I see the slowness more when I peek at VCF when I'm on break or eating lunch at my desk.

Hope that helps.

EDIT: PS, I used YSlow to help with some of the cache tuning, and image optimization, on my site.
 
This is the issue I know people are experiencing issues, I need to map when exactly.
FYI just now it took about 20 seconds to load this thread when I called it from my email link.
b
 
1) Every couple of days, and Apr. 27th @ 10:28AM EST -0500 GMT (today)
2) Takes about a minute or two to open anything, and 30-45 seconds to open the Quick reply, post editor, etc.
3) Systems running Windows 98SE, XP, 7, 10, Linux, and Android (Samsung Galaxy S7 Edge) with up-to-date browsers on all of them (or out-of-date) - IE, Chrome, FireFox, and Edge (Win10 Pro) and updates installed
4) Verizon Wireless (Samsung Galaxy S7 Edge), Xfinity (home ISP) in SE Michigan
Sometimes it works, other times it sits there and eventually comes up.

Thank you guys for working on this issue and resolving every other issue with this site. Kudos to you!
 
Last edited:
please provide date and time exactly so that we can gather data and create a specific case that we can submit to the ISP for support purposes. For example

1. April 27th 10:47AM EST
2. Thread page load lagged for 20 seconds before loading
3. Using Windows Explorer 11
4. Verizon FIOS
 
Ok, two actual data points, when replying to the Sun Ultra 5/10 thread with information about HVD versus SE SCSI:
1.) 4/27/2018 10:46AM (post: http://www.vcfed.org/forum/showthread.php?62967-Sun-Ultra-5-Linux-install-issues&p=512885#post512885 ).
2.) 45 seconds between clicking 'Post Quick Reply' and page reload after reply posted.
3.) Dell Precision M6700, Core i7-3740QM, Firefox 52ESR on CentOS 7.
4.) MCNC, 1Gb/s, Cisco 7609 edge router, North Carolina (multiple 100Gb/s uplinks to the Internet in Atlanta, GA; Charlotte, NC; Raleigh, NC; and Winston-Salem, NC.)

Data point number two:
1.) Reloading this thread, 4/27/2018 10:47 AM. Firefox's rotating circles in the tab had two distinct phases: the counter-clockwise rotating dot on a circle for 15 seconds, then the clockwise-rotating blue circle for about 20 seconds.
2 through 4: same as previous.

EDIT:
Data point number 3:
1.) Reloading this thread, 4/27/2018 11:15AM. Reload was nearly instantaneous, no rotating circles.
2 through 4: same as previous.
 
Last edited:
Ok 4-27-18 12:11pm EST, I was replying to a tread http://www.vcfed.org/forum/showthread.php?61569-Juko-D16-X-can-the-bios-be-upgraded
I hit the quick reply button. It took about 17 seconds and the screen changed to the Full post reply screen but my information was still populated in the text field. I hit submit and it showed my posting was duplicated.

I was replying to the thread the same way I do to all replies, nothing abnormal and I didnt hit the wrong button.
 
The running theory at the moment is that the server itself is underpowered.

This can be easily verified by gathering metrics on the server. In all of the VCF, surely someone is a unix sysadmin by trade, yes? Even opening a shell window and checking the load average would be better than guessing...

If nobody on the VCF is a unix sysadmin, I can volunteer my time, contact me in a PM.
 
I tried a few browsers on my system.
Firefox ESR and Palemoon 27.9 are fine, Chromium hangs when doing trying to do a reply seems to be a problem specific to that. Tor is slow alround.
This is on a Celeron 2.4 running Devuan Jessy i386
My ISP is SPARK here in NZ


Edit: hitting the Post button caused a 34 second delay and a "Stay on this page/leave" prompt.
Just had this happen. Selected leave and post was saved.
Apr 28 08:43 NZ time
Firefox ESR
Celeron 2.4 running Devuan Jessy i386
SPARK
 
Last edited:
From today:

1. 15:14 CDT
2. Loading "New Posts" took 18 seconds (in the past, took only 1-2 seconds)
3. Windows 10 64-bit, Chrome
4. Chicago, USA, ISP: Wide Open West

1. 15:16 CDT
2. Clicking "go advanced" button to edit this thread took 12 seconds (in the past, took only 1-2 seconds)
3. Windows 10 64-bit, Chrome
4. Chicago, USA, ISP: Wide Open West

1. 15:18 CDT
2. Clicking "Save Changes" button took 19 seconds (in the past, took only 1-2 seconds)
3. Windows 10 64-bit, Chrome
4. Chicago, USA, ISP: Wide Open West

1. 22:25 CDT
2. Clicking a user handle then "View Forum Posts" took 28 seconds (in the past, took only 3-4 seconds)
3. Windows 10 64-bit, Chrome
4. Chicago, USA, ISP: Wide Open West
 
Last edited:
1. 4/28/18 12:17 AM

2. Going between forum sections and using the "reply to thread" button take exorbitant amounts of time. 30+ seconds, sometimes over a minute.

3. Various machines at various locations, mostly using either Windows 7 x64 and Fedora 26/27 using Firefox or Chrome

4. Texas, Spectrum Internet.

A large portion of the page load seems to be from Facebook web bugs, 938ms in this test:

https://tools.pingdom.com/#!/eHvmhk...Let-s-work-together-to-determine-some-details
 
I block all the outside stuff with plugins (uBlock+Adblocker) and it still takes 10-15 seconds to go between anything in the forum. `

I'd say you've probably got some packet loss or the host is underpowered... ping tests and 'top' should be able to root out which pretty quickly. Do you have
sudo console access at the host?

loss.jpg

I show some packet loss in the peering right before your host. Some Network Engineers will tell you they drop ICMP that shows up this way and it means nothing, but that's only when the node is congested and they don't want to try to fix it, or can't. It could be as stupid as a bad patch cable or mismatched duplex. This is an average over 10 minutes- I saw it dropping 10% at one point. Dropping 5% of traffic over 10 minutes is a problem.

It turns out the attachment is unreadable due to how the forum squashes it- you can see my test at http://rittwage.com/files/loss.png. You can also generate your own and do it with Pingplotter Pro (trial version is fine).
 
Last edited:
Took 25 seconds to load this post and 13 to get this reply box. I'm pretty certain the ISP issue is localized to this server host. The latency I experience varies wildly throughout the day.

I host a lot of sites. If you need a second pair of eyes on things, let me know. I've offered to pay for upgraded hosting for this site and that still stands.
 
I show some packet loss in the peering right before your host. Some Network Engineers will tell you they drop ICMP that shows up this way and it means nothing, but that's only when the node is congested and they don't want to try to fix it, or can't. It could be as stupid as a bad patch cable or mismatched duplex. This is an average over 10 minutes- I saw it dropping 10% at one point. Dropping 5% of traffic over 10 minutes is a problem.

I have issues with that node at my location as well.

Code:
  1    <1 ms    <1 ms    <1 ms  192.168.1.1
  2    32 ms    24 ms    20 ms  cpe-70-112-96-1.austin.res.rr.com [70.112.96.1]
  3    42 ms    33 ms    30 ms  tge0-0-0.kyletx0502h.texas.rr.com [66.68.2.65]
  4    24 ms    20 ms    18 ms  agg26.ausxtxir02r.texas.rr.com [24.175.43.132]
  5    36 ms    24 ms    21 ms  agg22.hstqtxl301r.texas.rr.com [24.175.41.48]
  6    28 ms    23 ms    30 ms  bu-ether16.hstqtx0209w-bcr00.tbone.rr.com [66.109.6.108]
  7    39 ms    25 ms    31 ms  107.14.19.49
  8    31 ms    26 ms    27 ms  0.ae0.pr1.dfw10.tbone.rr.com [107.14.17.232]
  9    28 ms    32 ms    24 ms  ix-ae-23-0.tcore2.dt8-dallas.as6453.net [66.110.57.97]
 10    26 ms    30 ms    26 ms  ae6.cr10-dal3.ip4.gtt.net [199.229.230.85]
 11    58 ms    63 ms    58 ms  et-8-3-0.cr2-lax2.ip4.gtt.net [89.149.143.53]
 12    161 ms     *       122 ms  fe0-pronethosting.cust.lax02.packetexchange.net [216.193.192.162]
 13    62 ms    55 ms    58 ms  104.192.220.12

I don't get as much packet loss, but the response times are all over the place and usually really high.
 
Back
Top