Buttle Butkus's questions

Advice I've found on the internet recommends calculating your Apache MaxClients based on how much RAM your average Apache process takes up. Just glancing at top, for my site that seems to be around 650KB. I have 32GB of RAM, with MySQL using about 20GB. If I gave Apache just half the remaining, 6GB, then I could supposedly handle ~9231 Apache processes.

But each Apache process spawns a PHP process, doesn't it? And my average PHP process is about 40MB. If I account for that, I get about 147 processes. Vastly different numbers. Can anyone shed any light on this question?

Something odd is happening on my server. The Apache log shows lots of requests from the server's own IP address. I do not suspect the server is compromised. I suspect some PHP code is malfunctioning but I don't know how to trace it.

I had to block the server from itself in htaccess because too many sessions were being generated, but the requests are still occurring and resulting in access denied errors.

I need to find out what is generating the requests. Below I have pasted a 4 lines from the Apache error log, all of which were generated within the same exact second (of a total of 24 generated in that second!). I suspect that somewhere in a PHP script there is something causing this. How can I find out the source of these requests?

Note that the IP that shows up is not the localhost IP, but the server's external IP. That may be significant.

[Mon Sep 21 17:35:21.056244 2015] [access_compat:error] [pid 13385] [client 64.123.123.123:44407] AH01797: client denied by server configuration: /home/user1/public_html/403.shtml
[Mon Sep 21 17:35:21.065225 2015] [access_compat:error] [pid 13406] [client 64.123.123.123:44408] AH01797: client denied by server configuration: /home/user1/public_html/image/catalog/somepath/ABC-1234XYZ.jpg
[Mon Sep 21 17:35:21.065270 2015] [access_compat:error] [pid 13406] [client 64.123.123.123:44408] AH01797: client denied by server configuration: /home/user1/public_html/403.shtml
[Mon Sep 21 17:35:21.066066 2015] [access_compat:error] [pid 13155] [client 64.123.123.123:44409] AH01797: client denied by server configuration: /home/user1/public_html/image/catalog/someotherpath/DEF-AAA999Z.jpg

By the way, since August 23rd there are 535,263 lines in my Apache error log where my own server's IP is the client. This is a huge problem.

I've been adding them to /etc/basrhc, and that works most of the time.

But I noticed that when I su to another user, I lose all my aliases.

It turns out that /etc/bashrc is only loaded when the terminal is opened, not on login. And it seems that /etc/profile is loaded on login. So, is that where I should really have put all those aliases?

Ubuntu precise (12.04.1 LTS)

I'm rather new to PEAR.

I installed PEAR. Then, using pear I installed phpdoc.

It seems to be working great except for the graphing functions.

I ran this command:

/var/www/site5 $ phpdoc -f models/classes.php -t ./docs/classes

Collecting files .. OK
Initializing parser .. OK
Parsing files
Parsing /var/www/site5/models/classes.php
Storing cache in "/var/www/site5/docs/classes" .. OK
Load cache                                                         ..    0.026s
Preparing template "clean"                                         ..    0.069s
Preparing 15 transformations                                       ..    0.000s
Build "elements" index                                             ..    0.017s
Replace textual FQCNs with object aliases                          ..    0.151s
Build "packages" index                                             ..    0.015s
Collect all markers embedded in tags                               ..    0.015s
Build "namespaces" index and add namespaces to "elements"          ..    0.004s
Transform analyzed project into artifacts                          .. Unable to
find the `dot` command of the GraphViz package. Is GraphViz correctly installed
and present in your path?  12.465s
Analyze results and write report to log                            ..    0.004s
$

I realized that in my apache virtual host for this site I had this line:

    php_value include_path ".:/var/www/site5/includes"

And so I thought maybe that was preventing inclusion of other directories... ?

So I tried changing the line to this:

    php_value include_path ".:/var/www/site5/includes:/usr/lib/php:/usr/share/php"

That didn't work either, so I finally commented out the line, but still the same error.

In case this helps, inside of /usr/share, I ran this command:

/usr/share$ find -name "*GraphViz*"
./php/phpDocumentor/vendor/phpdocumentor/graphviz/src/phpDocumentor/GraphViz
./php/phpDocumentor/vendor/phpdocumentor/graphviz/tests/phpDocumentor/GraphViz
./php/Image/GraphViz.php
./php/test/Image_GraphViz
./php/data/phpDocumentor/features/generate-documentation/graphs/GenerateClassDia
gramUsingGraphViz.feature
/usr/share$

I don't see why this is not working. Thanks for your help.

I was confused about why my new server had CentOS 5.9 installed on it, when CentOS 6.3 is available.

But CentOS 5.9, it turns out, is actually the latest release.

Why are 5.x versions being released when they are a dead end that will lose full support barely 1 year from now?

Is there any reason to install CentOS 5.9 on a new server instead of a 6.x? Is there any reason to avoid 6.x? I assumed that 6 > 5.

There were other devs logged in as root that I wanted to kick out.

So I issued

pkill -KILL -u root

Then I realized that I probably killed all kinds of root processes.

Have I damaged my system? I can't reach it through SSH now.

UPDATE: Web server is still running. But I can't connect by SSH now. I have no idea what I've done.

Ubuntu server 12. Apache, MySQL, PHP

I've only ever used Apache. I know how to set up the Virtual Hosts and all that. But I have no idea how to add NGINX to the mix and not ruin everything.

I'm playing with a testing server, so it's ok if something goes down temporarily. What I would like to do is have ONE (or a few) of the testing sites on the server use NGINX while the others still use Apache. Is that even a thing? Can I use both on one server?

If not, then what I'd like to do is be able to switch between using Apache and NGINX serverwide. Could I install NGINX and easily switch between the two?

e.g. could I just stop Apache, start NGINX and have everything work smoothly (assuming NGINX is configured properly)?

I'm about to install nginx and I just want to make sure it doesn't break any existing configurations.

I tried adding lines to my.cnf

[mysql]
infile = 1
[mysqld]
infile = 1

SHOW VARIABLES LIKE "%infile" displays true in mysql.

But LOAD DATA does not work. I tried it in a PHP script and in phpMyAdmin.

It only works if I start mysql from the command line with the --allow-infile flag

mysql --local-infile -u root -p db_name

But this did not solve the PHP problem.

This is NOT your standard "how do permissions work" question!

I'm thinking the answer will probably involve ACL, but I don't know how exactly.

I've messed around with standard permissions, guid, sticky bit, etc. Doesn't work.

What I want: User will upload files. User will have ability to change permissions on files to allow execution. But user will not be able to change the contents of file once it's created. And user will not be able to delete the file.

Please help! I've been tearing my hair out for hours!

EDIT:

Thanks for the answers but so far they don't seem to address how to have this work automatically for newly-created files.

Possible solution:

find -mtime -1 -exec chattr +i '{}' \+

2nd possible solution:

inotifywait -m -e create --format %f .

Now I just need to figure out to pipe that to chattr.

I have some tar.gz files that total many gigabytes on a CentOS system. Most of the tar.gz files are actually pretty small, but the ones with images are large. One is 7.7G, another is about 4G, and a couple around 1G.

I have unpacked the files once already and now I want a second copy of all those files.

I assumed that copying the unpacked files would be faster than re-unpacking them. But I started running cp -R about 10 minutes ago and so far less than 500M is copied. I feel certain that the unpacking process was faster.

Am I right?

And if so, why? It doesn't seem to make sense that unpacking would be faster than simply duplicating existing structures.

One of my questions on here has a lot of views and I have yet to select the best answer. Unfortunately, I cannot remember how I solved my problem.

The problem involved installing some stuff from yum. So I did:

history | grep yum

But I got no results, probably because I ran those commands many months ago, some time around January I think...

Is there any way to dig up that old history?

CentOS system.

Summary: changed memory_limit in master and local php.ini and yet no change in the local value for a particular virtual host.

Trying to improve performance, I set the memory_limit to 1024M in /etc/php.ini

phpinfo() shows Master and Local values for other virtual hosts on the server as 1024M. Changing the value in /etc/php.ini changes all values, except one.

One site is stuck with a local value of 256M.

I thought I found the problem: there is a php.ini file (which I didn't know about) in that site's root, and it had

memory_limit = 256M

I changed it to 1024M. Problem solved? No. And now I don't know where to look. Obviously, I've restarted apache (/etc/init.d/httpd restart), and that usually does the trick.

I also turned off APC cache, though I don't think it would cache ini files.

And finally, I tried adding this to the virtual host in httpd.conf:

php_value memory_limit 536870912

(yes, that would be 512MB) And that had no effect whatsoever.

What else could be the problem?

Thanks.

I'm not sure if this is a question for serverfault or stackoverflow but it seems like it has a lot to do with server config.

We have a server in chicago and the server's clock is on chicago time. But since the business is located in California, it would seem to make sense to use pacific time.

What happens when server time is Chicago, and php.ini directive date.timezone is set to "America/Los_Angeles"? How will that affect logs written to mysql, error logs, etc?

I've looked at the Apache error log and, as I expected, the php directive does not affect it. Times are all servertime.

Thanks.

I installed php 5.3 and some modules using

yum install php53-{module-name}

But mcrypt did not work.

I searched like this:

yum list php* | grep mcry

The only package that came up was

php-mcrypt.x86-64         5.1.6-15.e15.centos.1    extras

I have the same exact problem with mhash. The results for php are even more confusing.

I know this is probably a dumb question, but any help would be appreciated. I am used to Ubuntu's simple "apt-get install" and everything works. CentOS seems to be a lot more detail oriented.

Thanks.

EDIT: Here's what I tried:

yum install php53-mcrypt

Result:

No package php53-mcrypt available

And trying this:

yum install php-mcrypt

Results in:

Error: php53-common conflicts with php-common

Because php-mcrypt tries to get php-common as a dependency (?) and I am trying to use php 5.3 rather than the default php 5.1 on CentOS 5.7.

I'm working on a fresh CentOS server and am extremely annoyed to find out that yum does not work as smoothly and brilliantly as "apt-get install" does on Ubuntu.

Installing a LAMP stack, I ran the following:

$ yum install php
$ yum install mysql

Furthermore:

$ yum install mysql-server
$ yum install mysql-devel
$ yum install php-mysql

Now running:

$ rpm -qa | grep php

...shows that I have all kinds of php 5.1.6 packages installed and I need 5.2 minimum. In addition, I see that it installed mysql 5.0.77, even though the current release is 5.5.20. But hopefully mysql 5.0 will work. PHP 5.1 will not.

So the real question(s) is(are):

How do I fix my PHP to make it the latest stable release (5.3.9 according to PHP.net)? How do I make my CentOS machine work more like Ubuntu so I don't have to deal with this again?

Thanks very much.

EDIT: CentOS 5.7 was installed by a hosting company last week, so I don't know why they would have installed an older version of CentOS, rather than the most recent. I need PHP 3.2 at least, so I appreciate the suggestions on how to do that. I'm not very familiar with packages and repositories, but it appears that anyone can start a repository and so I'm concerned about installing things from some random repository that could infect the server.

Our old server has it in /home/user/public_html/ but it seems like it makes more sense in /var/www/.

The old server had cpanel and I suppose that as part of that setup, it separates virtual servers into their own folders in /home/ and gives them each a web root. All the files in /home/user/ were also owned by user, rather than root.

But we don't have sub accounts and I'm used to creating various virtual hosts in /var/www/ (from my Ubuntu experience), so is there any downside to just using /var/www/ in CentOS? I would then have all the files owned by a group that would include me and the other developer.

I have MySQL installed on an Ubuntu machine.

I added this line in /etc/mysql/my.cnf

group_concat_max_len=15360

But there is no effect. Every time I restart mysql, the value is set to 1,024. I have to manually run

SET GLOBAL group_concat_max_len=15360

...every time I start up mysql.

Why is my.cnf not working the way I thought it should? Thank you

Looking at Xeon E3-1230 system with 8 GB DDR3 RAM.

I've read (somewhere) that Cent OS 64 Bit has problems and/or that the 32 bit version is "better".

Also, I know that with windows you need 64-bit in order to use higher amounts of RAM, but I don't know if that applies here.

Thanks

I'm specifically looking at the Standard Large On-Demand instance which has 850 GB of instance storage.

What I really don't understand is why have 850 GB of storage on the instance if that data will disappear as soon as the instance is shut down and/or destroyed.

It seems like EBS is the standard way of having permanent disk space for instances, including root volumes. You can stripe them. You can back them up in different availability zones and/or S3 and/or off of Amazon's system completely.

Why keep anything at all on the instance storage if you would have to back it up frequently off-instance (EBS/S3?) to avoid losing it all? Is it a speed advantage even beyond what striping EBS volumes would give you?

It seems like the best configuration for safety and ease of setup would be to just not use the instance storage at all, and to have instead striped EBS volumes with backup to S3 or off site.

Am I right, or is there a good reason for using that 850 GB of instance storage?

Thank you

My main virtual machine crashed today. More specifically, the Windows7 host crashed while the VM was transferring files (very slowly) to a flash drive. I'm running Ubuntu 10.04 LTS - Lucid Lynx in VMware (not sure if it is VMware Player, esx, or what. I downloaded it for free about a year ago) as a testing webserver. Now the VM will not boot past the 1st 2-3 lines. Then it does nothing at all. I waited about 30 minutes one time.

Usually it says this:

fsck from util-linux-ng 2.17.2
/dev/sda1: clean, 652189/6373376 files, 14459048/25464832 blocks

One time it said this:

fsck from util-linux-ng 2.17.2
/dev/sda1 has been mounted 27 times without being checked, check forced.
/dev/sda1: clean, 652189/6373376 files, 14459049/25464832 blocks

[Note: aside from the extra line, the only difference is the number of blocks in the numerator]

I have a bunch of custom php code on there. several mysql databases, and I've made lots of tweaks to the system that I would hate to lose.

To get into Recovery mode, I tried pressing and holding the -shift- key while booting up the VM, to no effect. I've also tried pressing -escape- a bunch of times. I've tried this on working VMs and it doesn't work there either. The only install where I get that option is the one I just did an hour ago, and it comes up automatically, without any key presses.

As I mentioned, the host system crashed while the VM was transferring files to a flash drive. The flash drive is special because I created an entry for it in fstab and it's mounted on every boot. The flash drive still works, by the way.

Does anyone know how I can fix this? One positive is that I now have access to the disk through another VM and all the files seem to be fine. Are there any particular files I should look at or modify to fix this?