Hi all,

I got a new RS server yesterday, thought it would be cake to setup and leave alone. But, it's REALLY slow for the first access (I've tried we access from both east and west coast) via the web, but then fast after that.

I got the server to just run phpadsnew. For example, the first time I go to the login page, it's REALLY slow (5-10 secs to pull up on fast connection). After that, it's fast - UNLESS I don't use the browser for 1 minute or so, then it's slow again. If I keep clicking around, it's fast, but if I let it sit for a while, it's DOG SLOW for the first access, then fast again.

It's not slow because of a server side issue (like big queries, or load) - it's does the same thing when I try to access simple static files.

I probably set something up wrong in CPANEL (it feels like the browser has to look up the server each time). I'm accessing it via the IP, not a domain name (I'm not planning on assigning it a domain name).

I'm officially out of my league on all this, so I defer to the great minds of WHT.

Could be the NIC, submit a trouble ticket.

I hate opening tickets at RS - I have one open from 22 hours ago that hasn't been worked on yet. I guess I will have to though.

Was hoping someone knew of a way I could have screwed it up (I get a popup saying I need to add an A entry all the time in WHM).

the only thing I think it would be is reverse dns look ups and the fact that you have no A entry for your hostname. But it seems its mostly a hardware issue.

I too would suspect the nic. RS is normally fast on supported issues (hardware/network).

I believe the problem was due to the /etc/resolv.conf file pointing to the sometimes-down CPanel nameservers that are on a completely different network. The suggestion was to change them to point to Rackshack's nameservers.

You guys were right, they responded very quickly.

Dear Customer,

I would suggest replacing the contents of your /etc/resolv.conf with the
following:
--- cut here ---
search rackshack.net
nameserver 207.218.192.38
nameserver 207.218.192.39
--- cut here ---

Your resolv.conf file is currently set up to use some of CPanels name servers, which are far away network-wise. This is likely the cause of the pauses. If this change doesnt fix the problem, feel free to re-open this ticket and add attempt to describe the problem in more detail.

Have a nice day.


About 1-2 hour turnaround, not bad.

Originally posted by jbridge21
I believe the problem was due to the /etc/resolv.conf file pointing to the sometimes-down CPanel nameservers that are on a completely different network. The suggestion was to change them to point to Rackshack's nameservers.

Thanks Jeffrey - pretty good for such a "newbie"!

OK, it's gone from slow to REALLY SLOW.

Just normal calls to static JPG files are PAINFULLY slow, like 5-10 seconds.

After they're loaded, hitting refresh takes FOREVER to re-load them (100 mb connection here).

Anyone know what could be causing this?

Sounds like a NIC card issue. Do this over SSH and post the results here. If it's slow, I bet you it's the NIC card.

wget ftp://ftp.uu.net/uumap.tar.Z

- Matt

Thanks Matt. It wasn't that slow,about the same as other RS servers I have (that work OK)

Here's the goofy server:
17:04:51 (658.58 KB/s) - `uumap.tar.Z' saved [8520221]

Here's two others:
17:19:18 (599.25 KB/s) - `uumap.tar.Z' saved [8520221]
17:31:07 (1.08 MB/s) - `uumap.tar.Z' saved [8520221]


Here's one from nocster, definately a lot faster:
18:04:27 (3.15 MB/s) - `uumap.tar.Z' saved [8520221]

THe nocster server is loaded much higher than the other RS servers, if that matters.

And, to add insult to injury, Rackshack just closed the ticket, without even looking at the problem - claiming it was the fault of the server that has the link to the files on the Rackshack server, which is not true (all he would have to do is right click on the file to see where it came from).

Everyone, click here - see how long this 6k image at RACKSHACK takes to load:

http://64.246.56.56/graphics/products/ufc40_small.jpg

Now compare to a LARGER image hosted somewhere else: http://www.sherdog.com/graphics/products/ufc40_large.jpg

Big difference, huh? This new Dual Xeon server has been slower than my original Cobalt RAQ 4i.

Even http://64.246.56.56/graphics/ took pretty long to lead for me - 6 seconds on a fast cable modem. Not sure what the problem could be, but it's definitly slow.

- Matt

well yes that is very slow

Microsoft Windows XP [Version 5.1.2600]
(C) Copyright 1985-2001 Microsoft Corp.

C:\Documents and Settings\Loren>ping 64.246.56.56

Pinging 64.246.56.56 with 32 bytes of data:

Reply from 64.246.56.56: bytes=32 time=65ms TTL=48
Reply from 64.246.56.56: bytes=32 time=63ms TTL=48
Reply from 64.246.56.56: bytes=32 time=63ms TTL=48
Reply from 64.246.56.56: bytes=32 time=62ms TTL=48

Ping statistics for 64.246.56.56:
Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 62ms, Maximum = 65ms, Average = 63ms

C:\Documents and Settings\Loren>tracert 64.246.56.56

Tracing route to rs-64-246-56-56.ev1.net [64.246.56.56]
over a maximum of 30 hops:

1 8 ms 17 ms 23 ms 10.49.112.1
2 9 ms 9 ms 10 ms c68.113.240.1.ona.wi.charter.com [68.113.240.1]

3 16 ms 15 ms 14 ms 24-216-136-121.charter.com [24.216.136.121]
4 14 ms 14 ms 33 ms 172.18.1.1
5 19 ms 20 ms 19 ms 172.18.98.3
6 19 ms 16 ms 18 ms 172.18.137.13
7 33 ms 33 ms 34 ms 12.124.129.81
8 31 ms 62 ms 33 ms gbr5-p40.sl9mo.ip.att.net [12.123.25.26]
9 35 ms 35 ms 38 ms tbr2-p013501.sl9mo.ip.att.net [12.122.11.121]
10 42 ms 44 ms 41 ms tbr2-p013801.cgcil.ip.att.net [12.122.10.45]
11 62 ms 40 ms 38 ms ggr2-p390.cgcil.ip.att.net [12.123.6.37]
12 43 ms 77 ms 41 ms att-gw.sf.prserv.net [192.205.32.194]
13 43 ms 44 ms 44 ms p16-1-0-0.r01.chcgil06.us.bb.verio.net [129.250.
5.77]
14 56 ms 58 ms 56 ms p16-0-1-1.r20.dllstx01.us.bb.verio.net [129.250.
5.85]
15 63 ms 62 ms 63 ms p16-6-0-0.r02.hstntx01.us.bb.verio.net [129.250.
4.25]
16 60 ms 64 ms 63 ms ge-0.ev1.hstntx01.us.bb.verio.net [129.250.10.66
]
17 102 ms 63 ms 63 ms 207.218.245.41
18 64 ms 62 ms 60 ms rs-64-246-56-56.ev1.net [64.246.56.56]

Trace complete.

C:\Documents and Settings\Loren> Put in a trouble ticket to rackshack for the gateway / router you sit on.. Seeing for me that is were the highest respons came from. And yes it could be a nic card problem also.

Good luck,
Loren

Garrett -

Looks like an application layer problem at first glance - L4 (tcp setup) looks pretty fast.

From my workstation:

1 2003-06-16 17:51:45.6959 207.218.xx.xx -> 64.246.56.56 TCP 2220 > 80 [SYN] Seq=3048204041 Ack=0 Win=16384 Len=0
2 2003-06-16 17:51:45.6960 64.246.56.56 -> 207.218.xx.xx TCP 80 > 2220 [SYN, ACK] Seq=3981463160 Ack=3048204042 Win=5840 Len=0
3 2003-06-16 17:51:45.7019 207.218.xx.xx -> 64.246.56.56 TCP 2220 > 80 [ACK] Seq=3048204042 Ack=3981463161 Win=17520 Len=0
4 2003-06-16 17:51:45.7041 207.218.xx.xx -> 64.246.56.56 HTTP GET /graphics/products/ufc40_small.jpg HTTP/1.1
5 2003-06-16 17:51:45.7046 64.246.56.56 -> 207.218.xx.xx TCP 80 > 2220 [ACK] Seq=3981463161 Ack=3048204356 Win=6432 Len=0
6 2003-06-16 17:51:46.5664 64.246.56.56 -> 207.218.xx.xx HTTP HTTP/1.1 200 OK
7 2003-06-16 17:51:46.5665 64.246.56.56 -> 207.218.xx.xx HTTP Continuation
8 2003-06-16 17:51:46.5904 207.218.xx.xx -> 64.246.56.56 TCP 2220 > 80 [ACK] Seq=3048204356 Ack=3981466081 Win=17520 Len=0
9 2003-06-16 17:51:46.5911 64.246.56.56 -> 207.218.xx.xx HTTP Continuation
10 2003-06-16 17:51:46.5912 64.246.56.56 -> 207.218.xx.xx HTTP Continuation
11 2003-06-16 17:51:46.5913 64.246.56.56 -> 207.218.xx.xx HTTP Continuation
12 2003-06-16 17:51:46.6150 207.218.xx.xx -> 64.246.56.56 TCP 2220 > 80 [ACK] Seq=3048204356 Ack=3981469001 Win=17520 Len=0
13 2003-06-16 17:51:46.6210 207.218.xx.xx -> 64.246.56.56 TCP 2220 > 80 [ACK] Seq=3048204356 Ack=3981470297 Win=16225 Len=0
14 2003-06-16 17:51:46.6315 207.218.xx.xx -> 64.246.56.56 TCP 2220 > 80 [RST] Seq=3048204356 Ack=3981470297 Win=0 Len=0

Notice L4 is extrememy fast. L7 appears to take around 3/4 of 1 second to respond.


Another more extreme example:

47 2003-06-16 17:51:59.0612 207.218.xx.xx -> 64.246.56.56 TCP 2224 > 80 [SYN] Seq=3051759287 Ack=0 Win=16384 Len=0
48 2003-06-16 17:51:59.0615 64.246.56.56 -> 207.218.xx.xx TCP 80 > 2224 [SYN, ACK] Seq=3986379655 Ack=3051759288 Win=5840 Len=0
49 2003-06-16 17:51:59.0674 207.218.xx.xx -> 64.246.56.56 TCP 2224 > 80 [ACK] Seq=3051759288 Ack=3986379656 Win=17520 Len=0
50 2003-06-16 17:51:59.0702 207.218.xx.xx -> 64.246.56.56 HTTP GET /graphics/products/ufc40_small.jpg HTTP/1.1
51 2003-06-16 17:51:59.0703 64.246.56.56 -> 207.218.xx.xx TCP 80 > 2224 [ACK] Seq=3986379656 Ack=3051759691 Win=6432 Len=0
52 2003-06-16 17:52:02.6007 64.246.56.56 -> 207.218.xx.xx HTTP HTTP/1.1 304 Not Modified
53 2003-06-16 17:52:02.6035 64.246.56.56 -> 207.218.xx.xx TCP 80 > 2224 [FIN, ACK] Seq=3986379936 Ack=3051759691 Win=6432 Len=0
54 2003-06-16 17:52:02.6184 207.218.xx.xx -> 64.246.56.56 TCP 2224 > 80 [ACK] Seq=3051759691 Ack=3986379937 Win=17240 Len=0
55 2003-06-16 17:52:02.6187 207.218.xx.xx -> 64.246.56.56 TCP 2224 > 80 [RST] Seq=3051759691 Ack=2156215650 Win=0 Len=0

Notice L4 took a fraction of a second, but the application took more than 2 seconds to respond.


The connectivity was perfect:

192.23#ping 64.246.56.56 co 40000
Sending 40000, 16-byte ICMP Echo to 64.246.56.56, timeout 5000 msec, TTL 64
Success rate is 100 percent (40000/40000), round-trip min/avg/max=0/0/6 ms.


I would guess the problem is a configuration issue on the server within the web application...

Sincerely,

Randy Williams, CTO
Everyones Internet, Inc.
rackshack.net

Hi Randy,

Thanks for the well-researched response. I guess it shows there's nothing wrong with the hardware (I think that's what you're saying).

What is an application layer?

All I did was get the box from Rackshack, and it's always been slow from day one.

If there's something wrong with the way the server is delivering files, then it was setup wrong, I've done nothing except run the upgrade scripts in cpanel.

What should I do, since I got a box with a bad application layer?

Garrett -

The application is Apache in this case.

You might consider ordering a restore. The evidence suggests there is nothing wrong with the hardware or network, and if that is the case the software installation/configuration becomes a prime suspect. (You might also present your case to customer service for a free restore.)

Sincerely,

Randy Williams, CTO
Everyones Internet, Inc.
rackshack.net

Try an apache rebuild first (if you already have data on the machine) - SSH as root, and do this:

/scripts/easyapache

Choose the option you'd like, sit back, and wait 10-20 minutes for Apache to recompile. If it doesn't work then, get a system restore.

- Matt

Thanks Randy. I edited the trouble ticket and asked for a restore.

Since I've had the server maybe a week now, and it's had trouble since day one (see this thread) I do think Rackshack should handle the restore for free.

Originally posted by mainarea
Try an apache rebuild first (if you already have data on the machine) - SSH as root, and do this:

/scripts/easyapache

Choose the option you'd like, sit back, and wait 10-20 minutes for Apache to recompile. If it doesn't work then, get a system restore.

- Matt

OK, it's compiling (option 1) - we'll put this racehorse through it's paces.

Originally posted by Garrett
Thanks Randy. I edited the trouble ticket and asked for a restore.

Since I've had the server maybe a week now, and it's had trouble since day one (see this thread) I do think Rackshack should handle the restore for free.

Garrett no problem on the free restore, please include this thread URL with your restore and drop me an email to ensure everything gets taken care of.

Thank you for your patience regarding the matter.

Fast 20 minutes :) Still painfully slow. After it restarted apache, it keeps returning this line:

Scanning suexec_log...Done

Over, and over, and over...

Thanks Mario, Email sent. I've already included this threadID in the trouble ticket. I appreciate the fast response Rackshack has shown today, thanks.



Originally posted by EV1-OSC
Garrett no problem on the free restore, please include this thread URL with your restore and drop me an email to ensure everything gets taken care of.

Thank you for your patience regarding the matter.

Scanning suexec_log...Done

Over, and over, and over...
Don't worry, that's supposed to happen for some odd reason.

- Matt

OK, so they restored my server (I think). I had copied all my files to the backup drive, but now that the server is back "live" I can't see the backup drive. I thought they would re-mount it or do whatever is needed, since it was SUPPOSED to be all setup as I paid a hefty setup fee.

Here's what I see:

Filesystem 1k-blocks Used Available Use% Mounted on
/dev/sda3 68437272 2061300 62899508 4% /
none 0 0 0 - /proc
usbdevfs 0 0 0 - /proc/bus/usb
/dev/sda1 101089 14315 81555 15% /boot
none 0 0 0 - /dev/pts
none 515280 0 515280 0% /dev/shm


Since the RS support ticket system is DOA (at least has been for the last 6 hours), can anyone suggest a good newb procedure for installing this SCSI drive so I can get up and running again?

Garrett,

I think you need to have someone who knows about linux servers to handle those issues for you.

Kev

I agree, but I'm broke, and already paid hundreds of $$$ for the host to set it up. I'm impatient, and looking to learn.



Originally posted by inteltechs
Garrett,

I think you need to have someone who knows about linux servers to handle those issues for you.

Kev

Try this,

mkdir /mnt
mount -t ext3 /dev/sdb1 /mnt

This is assuming your second hard drive is a scsi, and the partition is located at /dev/sdb with type ext3.

Peter

The other day i was checking up my server on Rackshack..I find it so slow..then searched for reasons...i guess 10mins..then find out
1. Bandwidth exceeded on my personal account by 20%
so then I change my settings and the child settings so it will probably move faster

yes...faster!