VPS server issues

[PhilH]

I have recently moved from a dedicated server to a VPS server from a different hosting company and I am having some problems.

The server has been up for 5 or 6 days now and it is still unreliable. I keep getting cpsrvd failures. I have set up some accounts, email, etc.. and on one account a number of databases to test Tiny Portal. This is all that I have done. No active websites other than the TP test site.

The support staff initially was very helpful but now I am feeling like I am getting the run around. I admit that my lack of experience has contributed to the problem at least in terms of communicating the problems and my inability to fix the performance problems myself.

I am now being told that "As for your concerns with our responses, please understand that every VPS we have is different". Question: Does this make sense?

Also, here is the most recent support ticket reply:

Firstly, thank you for you continued patience and assistance with these issues. With regards to your quoted response, I believe that Name is concerned that he would like to understand some more of the issues - so if we do create this parallel VPS we can understand any and all of the issues that arise and deal with them accordingly.  Given currently we still are trying to ascertain what causes the issues, it could be duplicated on the new VPS and also cause other problems that we have already corrected on your current.  I will discuss this further with him today.

1. cpsrvd keeps failing:
I have run a script on the VPS to improve the memory management on the server. It should help with the limits and the management of the memory should those limits be reached.  We have also submitted a request to enable us to monitor processes or requests that are causing memory spikes within the VPS.  Other monitors such as website or IP Port availability can certainly be done at this point, but we need a lower level monitor to watch the processes.

2 + 3. MySQL admin issues and phpmyadmin:
The MySQL admin could be one of the processes that is giving the issues to cpsrvd, as such we cannot prove this without the monitor that we have requested.  It could be that the process never executes or fails which renders either the blank screen or the non-execution.  The phpmyadmin I believe to be by design, cPanel manages the MySQL database and as such databases should be created via that interface. 

4.cpanel is unreliable especially when working in File Manager as in cpsrvd
failures:
The script I ran should also assist with this - I would like to monitor the VPS and see if some issues occur today. 

I would also like to perform a reboot of the VPS to ensure that the new settings for the memory take place, however I do not want to do this without your authorization. Could you either perform the reboot or let me know a suitable time to perform for you.

As for your concerns with our responses, please understand that every VPS we have is different - as such we need to understand what causes and what occurs when problems arise.  We are making every effort to resolve your issues, unfortunately sometimes these do take time. We deal with tickets when they come into our ticketing system, and as such we move through the list.  Name and I are covering this ticket and are both watching it to ensure that it gets the proper attention.  There is also investigatory stages that we need to go through to get to certain points of resolution.  I hope that clears things up in your mind with regards to how your ticket and your server are being handled by us here at support.  I have already extended your hosting with us by a month to show you our dedication and to ensure that you are not "paying" for these issues.

I am not looking to bash the host, simply looking for help with dealing with them. I went  them because of their reputation for good customer service and want to make things work out.

Any feedback/help would be appreciated.

[Christian A. Herrnboeck]

cpsrvd is the cPanel / WHM daemon... Go into your WHM, and add /scripts/upcp to the end of the URL. Wait till it completes.

Click on Updates -> System Software, let that complete
Click on Updates -> Server Software, let that complete.

Do a soft-reboot of the system, and your problems should be solved. If they are not, cPanel is running out of memory, which, in my experience is caused by too little memory allocated by the VPS software to your particular VPS.

2. MySQL is administered via cPanel, NOT phpMyAdmin. phpMyAdmin does not have permissons to create databases when used from within cPanel, you must do that through the MySQL link in the site's controlPanel.

3. I've never heard that cPanel is unreliable... 95% of all hosts out there use it... I have two servers running it, never had a problem with cPanel / File Manager... that's BS

4. How could each VPS be different? You install the VPS software on one server, then configure cPanel / WHM and related softwares on each. THe only difference is after you start installing software on it...

Finally, Post part (i.E one day's worth) of kernel logs... the answer to cpsrvd failing would most likely be contained in there.

Sincerely,
Christian A. Herrnbock

[PhilH]

Thanks for the quick reply...very helpful

cpsrvd is the cPanel / WHM daemon... Go into your WHM, and add /scripts/upcp to the end of the URL.

Can you expand on this a bit. Just a little more instruction.

Finally, Post part (i.E one day's worth) of kernel logs... the answer to cpsrvd failing would most likely be contained in there.

I will work on doing this.

[PhilH]

Quote
cpsrvd is the cPanel / WHM daemon... Go into your WHM, and add /scripts/upcp to the end of the URL.
Can you expand on this a bit. Just a little more instruction.

Ok, after a bit of thinking I got.

It got to:

Fetching version information from http://updates.cpanel.net/eximinstall/version (RELEASE)....Done
open3: fork failed: Cannot allocate memory at /scripts/cPScript/SafeRun.pm line 45
Fetching http://layer2.cpanel.net/installer/ (0)[email protected]%...91%...100%......Done

Then a dialogue box popped up, do you want to continue? contains nonsecure items, Do you want to continue?
I selected yes and then I got "page unavailable". At this point I started looking for the "cpsrvd failure" email in my inbox. Guess what? I have mail. Getting very tired of that email.

[PhilH]

Finally, Post part (i.E one day's worth) of kernel logs

current error_log contents

Out of memory!
Out of memory!
Content-type: text/html

Content-type: text/html

Content-type: text/html

Internal Error: "" Normal Connection Timeout (Probably closed browser on a HTTP/1.1 session) at cpsrvd.pl line 1460.
Internal Error: "" Normal Connection Timeout (Probably closed browser on a HTTP/1.1 session) at cpsrvd.pl line 1460.
Internal Error: "" Normal Connection Timeout (Probably closed browser on a HTTP/1.1 session) at cpsrvd.pl line 1460.
Child [29281]: exited with signal 13
Content-type: text/html

Content-type: text/html

Internal Error: "" Normal Connection Timeout (Probably closed browser on a HTTP/1.1 session) at cpsrvd.pl line 1460.
Content-type: text/html

Internal Error: "" Normal Connection Timeout (Probably closed browser on a HTTP/1.1 session) at cpsrvd.pl line 1460.
Internal Error: "" Normal Connection Timeout (Probably closed browser on a HTTP/1.1 session) at cpsrvd.pl line 1460.
Content-type: text/html

Internal Error: "" Normal Connection Timeout (Probably closed browser on a HTTP/1.1 session) at cpsrvd.pl line 1460.
Content-type: text/html

Internal Error: "" Normal Connection Timeout (Probably closed browser on a HTTP/1.1 session) at cpsrvd.pl line 1460.
Child [32003]: exited with signal 13
Content-type: text/html

Internal Error: "" Normal Connection Timeout (Probably closed browser on a HTTP/1.1 session) at cpsrvd.pl line 1460.
Content-type: text/html

Internal Error: "" Normal Connection Timeout (Probably closed browser on a HTTP/1.1 session) at cpsrvd.pl line 1460.
Child [3702]: exited with signal 13
Child [5814]: exited with signal 13
Child [9691]: exited with signal 13
Child [16289]: exited with signal 13

[Christian A. Herrnboeck]

Hi,

What I meant was, if your ip is 255.255.255.255, then go to:

https://255.255.255.255:2087/scripts/upcp

That'll update cPanel for you.

Ok... this confirms it; the problem is with memory allocation; under your system information, how much memory does it say you have? How about swap space, this is listed as the /tmp partition on the HDD.


-Christian

[[Unknown]]

Hmm... we had some problems with cPanel not loading on this server, actually.  Updating to the latest version in WHM and restarting the server has fixed it.

Can you upload the attached file and post a link to it, please?

-[Unknown]

[PhilH]

Can you upload the attached file and post a link to it, please?

Glad to comply except I am not certain where to upload to?
Two options: You could provide some instructions, which I would certainly appreciate or I can give you access to the machine via pm.

To all, thanks for the help

[[Unknown]]

Oh, I'm sorry.  The same directory as SMF is usually preferrable, but it should work even in a different directory.

-[Unknown]

[PhilH]

I figured it out, your post came in while I was typing this))

[[Unknown]]

Odd.  While the memory is used up (which is actually normal) you have plenty of swap space.  As much as it might be slow, nothing should be crashing for want of memory, I would think.  And your load averages are quite good.... I assume you're not using it in production quite yet?

Have you been able to get to WHM (or, do you even have WHM access) to update cPanel?

-[Unknown]

[PhilH]

yes on updating cpanel, done by the host and I also did it but got the ugly cpsrvd email. I do believe there was some improvement but I am convinced memory issues would be a recurrring problem with a VPS.

On a brighter note, the host has offered to move me to a dedicated server, a "developement box" for a heck of a price and no charge for the first month. I am once again impressed with their customer service. More about that later.

The present server will be up for a bit if you want to do some exploring [Unknown].

[PhilH]

On a brighter note, the host has offered to move me to a dedicated server, a "developement box" for a heck of a price and no charge for the first month. I am once again impressed with their customer service. More about that later.

[Unknown]   I probably should have brought this up in my first reply to you. I guess I just got so excited about having your help that it did not occur to me to bring it up. My apologies and no offense intended.

[LiroyvH]

To fix cpsrvd, al you have to do is run /scripts/upcp --force (execute that command on root shell, information based on that you have enough free space, etc.)

[PhilH]

Follow up:
Below is a sample of the QoS alert log in the Virtuozzo contol panel on my VPS. This time frame is after updating cpanel as desctibed above. The times are PDT. There are not any production sites on the server. Also I was not doing anything on the server myself. I have been told these yellow, red and black warnings are indicative of memory issue on a VPS.

Sep 09, 2005 03:13:43 AM Green zone  privvmpages
Sep 09, 2005 03:13:43 AM Red zone  numiptent
Sep 09, 2005 03:08:42 AM Black zone  privvmpages
Sep 09, 2005 03:08:42 AM Black zone  numiptent
Sep 09, 2005 02:03:55 AM Red zone  numiptent
Sep 08, 2005 04:48:34 PM Green zone  privvmpages
Sep 08, 2005 04:48:33 PM Yellow zone  privvmpages
Sep 08, 2005 04:39:04 PM Green zone  privvmpages
Sep 08, 2005 04:38:24 PM Yellow zone  privvmpages

Thought this might be of interest to someone who runs into similar problems.
Thanks to all for your help.