I've got a few machines where Apache acts really strange and curious if anyone has any suggestions. I'd love to figure this out so it can actually be deployed to a larger amount of machines and not just test instances.
- Basic Information
Apache 2.2.8 (Tried a few 2.2 versions)
PHP 5.2.6
suPHP based
Prefork Based
- Once a day at a random time Apache fails a request from remote monitoring. It comes back within a minute however is is inaccessible for that time. It sometimes gets picked up by 5 minute monitoring on the machine itself and it restarts the service obviously.
- PHP scripts fail to be killed at times resulting in memory being used. They need to be killed in order to go away.
Worker Based
- Apache can stay up forever it does not fail any requests
- PHP scripts do not get killed at a more frequent basis than in prefork. You need to `kill -9 pid` in order to get rid of the php processes.
I read about very few issues with 2.2 so I'm quite confused by this.
I'm using CPanel with Apache 2.2.x, php5 with suphp.
My server use multi processor:
Dual Quad Core E5405 with 8GB RAM. I should use mpm worker or prefork and what's best config for? (i hosted over 200 domains)
I tried to Google about my question, someone say mpm worker is better, someone say mpm prefork is better. I dont know what's exactly. Can you advise me something?
I am really not familiar with MPM prefork. Should apache use MPM prefork or is there a better alternative? If yes, what would that be and how to use that alternative?
I'm building a new server for a predominantly php5/mysql5 website and was wondering which version of apache to put on. I know there are some issues with PHP and Apache in MPM, but what's the performance comparison between Apache 1.3 and Apache 2 Prefork?
I've had a server running for a year or so and all of the sudden the server error log started outputting a very strange message that i've never seen before:
Code: 'groups' is not recognized as an internal or external command, operable program or batch file.
From what I understand this means apache is trying to run a program that doesn't exist, what could be triggering this?
I recompiled apache on my server but i never checked the "mpm prefork (recommended by apache)" option. So a few days ago I ask a server admin about if i needed that or not? I told him to go ahead and recompile it in there. So today I checked apache logs and got these
ar 12 20:13:51 host kernel: httpd[6806]: segfault at 000000000000000f rip 0000002a9a56287e rsp 0000007fbffff090 error 4 Mar 12 20:13:58 host kernel: httpd[6861]: segfault at 000000000000000f rip 0000002a9a56287e rsp 0000007fbffff0a0 error 4 Mar 12 20:14:03 host kernel: httpd[6804]: segfault at 000000000000000f rip 0000002a9a56287e rsp 0000007fbffff090 error 4 Mar 12 20:14:08 host kernel: httpd[6890]: segfault at 000000000000000f rip 0000002a9a56287e rsp 0000007fbffff0a0 error 4 Mar 12 20:14:12 host kernel: httpd[6776]: segfault at 000000000000000f rip 0000002a9a56287e rsp 0000007fbffff0a0 error 4 Mar 12 20:14:14 host kernel: httpd[6923]: segfault at 000000000000000f rip 0000002a9a56287e rsp 0000007fbffff090 error 4 Mar 12 20:14:57 host kernel: httpd[6955]: segfault at 000000000000000f rip 0000002a9a56287e rsp 0000007fbffff0a0 error 4
Is this related to the MPM prefork? Do you even need to check that box (mpm) to do settings like maxclient, maxchildprocess etc etc?
I did my own googling of the problem I some said it was because of php and/or hardware (ram). But that server admin guy says it's most likely the mpm prefork. Anyone have thoughts about this?
I'm using Plesk 8.1 on Server AMD Athon 3400, 1GB RAM. Something wrong on my server, but i don't now, how to fix it. Average load is normal, but by Task, for sometime are going up to over 250 and in this time can't anybody connect to server ( HTTP or FTP ).
By top command, i founded most of cpu press by user wwwrun, so run this command top -u wwwrun, here is result :
There are multiple occurrences of this at any one time, and the interesting thing is that it appears to be spoofing the source IP addresses - most are all different with few exceptions.
Has anyone else seen this and know of a solution? Normally I would simply use IP deny but given the addresses appear to be spoofed and too numerous it would be futile.. I thought if I programmed OSC to quit if it matched the keywords might be a decent solution, but so far I haven't had any luck
I searched google and this forum to see if I could find out anything with no luck at all, so I'm guessing this is fairly new.
I found a strange PHP file in a strange folder on a VPS I am using to host a few sites. I've looked through the logs but can't figure out how it got there and I've look at the code and can't make any sense of it. Can somebody take a look at the code and tell me what they think of it: .....
This month I just pruchase dedicated server, spec are AMD X2 with 1GB RAM.
On ssh, the memory result is: root@server1 [~]# free -m total used free shared buffers cached Mem: 883 836 47 0 163 397 -/+ buffers/cache: 275 608 Swap: 2047 0 2047
My question:
1. Why the total ram just 883MB? I think it should 1024Mb?
2. The server still empty, but why I see the total used memory is 836Mb?
I only have experience with cpanel vps and when my server empty it only use around 200MB RAM and around 400MB ram usage when my vps load with 30+ account.
A couple of days ago I came across www.just-ping.com site (it's a simple ping test site).
I tested my site avensen.com (IP: 72.232.147.154) with it, and got bad results like this one:
[url]
Santa Clara, U.S.A. Packets lost (20%) 50.6 51.9 52.8 Florida, U.S.A. Packets lost (80%) 45.6 45.6 45.7 Vancouver, Canada Packets lost (80%) 56.5 56.6 56.7 New York, U.S.A. Packets lost (20%) 50.7 57.2 61.5 Austin, U.S.A. Packets lost (60%) 9.5 9.6 9.9 Austin, U.S.A. Packets lost (90%) 9.4 9.4 9.4 Amsterdam, Netherlands Packets lost (60%) 121.6 122.4 123.3 Amsterdam1, Netherlands Packets lost (60%) 121.5 123.6 125.6 London, United Kingdom Packets lost (90%) 111.4 111.4 111.4 Sydney, Australia Packets lost (90%) 200.2 200.2 200.2 Stockholm, Sweden Packets lost (20%) 144.7 147.7 148.3 Cologne, Germany Packets lost (80%) 133.3 135.6 137.8 Madrid, Spain Packets lost (70%) 150.7 150.8 151.0 Paris, France Packets lost (60%) 128.4 132.5 135.5 Hong Kong, China Packets lost (30%) 196.1 196.4 196.8 Munchen, Germany Packets lost (60%) 131.7 131.8 132.0 Kraków, Poland Packets lost (70%) 196.3 198.5 200.2 Cagliari, Italy Packets lost (40%) 154.9 155.3 156.3 Melbourne, Australia Packets lost (50%) 199.6 205.5 208.2 Singapore, Singapore Packets lost (70%) 257.4 260.3 262.5
I'm trying to figure out if this is a network problem or a problem with my server. I don't get it, because there are no lost ICMP packets when I ping another hosts from my server, or when I ping my server from home PC.
And here is what server4sale support wrote:
Quote:
This is what we received from data center and will update you, when they get back to us.
"We apologize for the delay in responding to you. We are aware of an issue that involves our upstream provider, and we have opened a ticket with them to get the issue resolved ASAP. We have asked them to investigate this issue and attempt to isolate the cause. Once we have more information from them, we will update you here in this ticket.
In the meantime, if you note any changes (good or bad), please provide traceroutes BOTH "TO" your server, and "FROM" your server, as well as a 300 count ping summary. This request has been made by our upstream provider, as we will forward any additional pings and traceroutes we receive directly to them. Without the traceroutes both to and from the servers, the information will not be useful for their investigation.
We will provide you with updates thru this ticket as we receive information from our provider. If you have any additional questions, or need further assistance, feel free to contact us. We appreciate your patience, while we work with to resolve this issue."
Second message from support:
Quote:
The data center has informed that they have not yet received an update from their upstream provider as they used to inform them after performing changes.
However, for better investigation and providing the results more precisely to their upstream provider they have asked you to provide the latest:
Quote:
1) 300 ping results from your PC to server
2) Traceroute from your PC to Server and
3) Traceroute from Server to your PC
I'd really appreciate if you help me to get these results and isolate the problem.
IP of my server: 72.232.147.154
What's even stranger is that when I run a just-ping.com test over 72.232.147.174 IP (a machine in the same SAVVIS data center, I guess), I get "All OK results":
I’m running RHEL 3, Apache and Cpanel. When I ran: "netstat –an" I found this in the results:
tcp 0 0 11.11.111.229:49158 11.11.111.229:80 ESTABLISHED tcp 0 0 11.11.111.229:49578 11.11.111.229:80 ESTABLISHED
If I’m reading this right these two unprivileged ports are open and talking to my privileged http port 80. Does this seem right? Why would these two ports on my machine have a connection. All this attention was sparked by abnormal spikes in load. Now I’m getting paranoid that something may be off even though I’m clean when scanning for rootkits etc…
I'm very new to dedicated hosting, but not to server admin in general, and have come across what seems to be to be a problem.
I'm based in the UK, and the dedi I went with is in the US, i have several VPS in the US, and I can download to them pretty consistently from a UK based server at around 5MB/s... this is on a VPS.
The dedi I signed up to lease has a 100Mb card, and a fairly well known provider, and yet the connection I get to the UK is terrible. It fluctuates wildly between 200KB/s and 5MB/s, seemingly at random, for example, downloading a 100MB file, i'll start at 500KB/s initally, and within a few seconds it might be 3.5MB/s, this could then go either way, but i'll usually end up with an average of about 800KB/s - which really seems awfully slow.
The traceroutes appear fine, there's around 110ms on ping and that is consistant, and similar to the figure the VPS get.
I've been in touch with their support, and after trying the usual suspects - including swapping the NIC - they lost interest. I was actually very impressed with them up until this point so feel pretty let down.
Is this normal? I've honestly never seen a download vary so wildly in speed. Unfortunately i'm tied in on a 3 month contract otherwise i'd drop them in a heartbeat right now.
All accounst in my dedicated server start to show a very strange error_log with the following entries:
==== [04-Nov-2009 21:28:51] PHP Warning: PHP Startup: Unable to load dynamic library '/usr/local/lib/php/extensions/no-debug-non-zts-20060613/php_interbase.dll' - /usr/local/lib/php/extensions/no-debug-non-zts-20060613/php_interbase.dll: cannot open shared object file: No such file or directory in Unknown on line 0 ..... ====
Always when a php script is accessed, new entrie with this error above is created.
I dont understand because php script have not any relation with intebase or pgsql and my server have not this e db installed.
I have an issue I've never seen before, and hope someone here can shed some light on this odd problem.
Just brought a new server online running Apache 2.2.3. The DocumentRoot was originally set to /var/www/html but I changed it to /home/[sitename]/public_html. I've done hundreds of Apache configs before and this one is no different.
Anyhow, here's the issue:
In SSH (logged in as root), if I create a file (let's just say I do a vi index2.php, enter "test" and save), when I try to go to pull the file in my browser, I get a 403 permission denied error. However, if I create the file in /var/www/html then move it to /home/[sitename]/public_html, the file comes up fine in a browser. Same goes for JPEG images. If I wget a JPG from another site directly into /home/[sitename]/public_html, get a 403 permission denied. If I wget it into /var/www/html then move it, it loads just fine.
It's important to note that the file permissions, owner, and the actual file itself are all identical. Both directories are chowned and chmodded identically.
I run WHMCS as my billing system for my company and everything is working fine, except piping emails to WHMCS.
The people I am hosting WHMCS with are running PHP 4.4.8 and so I have the loaders for 4.4 uploaded to my account. When I try to send an email to an address that is set to pipe to WHMCS I recieve the following:
This message was created automatically by mail delivery software.
A message that you sent could not be delivered to one or more of its recipients. This is a permanent error. The following address(es) failed:
pipe to | /usr/local/bin/php -q /home/user/domains/dephnet.com/public_html/my/pipe/pipe.php generated by info@domain.com local delivery failed
The following text was generated during the delivery attempts:
------ pipe to | /usr/local/bin/php -q /home/user/domains/dephnet.com/public_html/my/pipe/pipe.php generated by info@domain.com ------
Site error: the file <b>/home/user/domains/dephnet.com/public_html/my/pipe/pipe.php</b> requires the ionCube PHP Loader ioncube_loader_fre_5.2.so to be installed by the site administrator.
Now why would the script want ioncube_loader_fre_5.2.so when the site works fine?
I just found hundreds of rubbish urls in awstats for a particular domain. Is this referrer spam or something more serious and can I do something about this?
Lately we have been getting log entries similar to the following from different IPs all over the US:
74.249.4.234 - - [03/Jun/2008:18:12:36 -0500] "GET / HTTP/1.1" 200 6205 "-" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1;1813)"
74.249.4.234 - - [03/Jun/2008:18:12:37 -0500] "GET /scripts/javascript.js HTTP/1.1" 200 9153 "-" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1;1813)"
74.249.4.234 - - [03/Jun/2008:18:12:37 -0500] "GET /scripts/overlib.js HTTP/1.1" 200 50733 "-" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1;1813)"
That is all there is to each hit.
Obviously, the default index.php file is being loaded and is calling the javascript files, but what we can't understand is why the CSS files and images are not being downloaded as well.
Any ideas on why this would be occurring?
Caching and text based browsing are unlikely scenarios due to the quantity and varied locations of the IPs.
I have one server with multiple websites hosted on it. It's powered with Cpanel / WHM.
When I run the command 'wget mydomain.com' from this server, I download a Cpanel / WHM landing page instead of the actual homepage from mydomain.com.
I can successfully load mydomain.com in a browser. I can also run the 'wget' command from my local computer and download the correct homepage.
When I run this command from the server mydomain.com is hosted on, Apache returns the home page for the default virtual host for my IP address (ie. cpanel landing page). I encounter this same problem when using CURL or PHP Sockets.
I am behind a firewall. Could this be causing the issue? Does anyone have any ideas how I could fix this?
This is a big problem as I have websites that need to use a web service from a domain hosted on the same machine.
Anyone is an expert for apf firewall here? I check my apf_log and saw it was automatically restarted every 12 hours. This is strange because the developer mode was disabled, and usually developer mode was only for 5 minutes and not for 12 hours.
Yes, the apf was restarted and function, but it's strange to see it restarted every 12 hours.
Using FreeBSD 6. These 2 lines in /etc/rc.conf resulted in not able to boot:
firewall_enable="NO" firewall_logging="YES"
The second rule had to be comment out otherwise it wouldn't boot. I started with "firewall_enable="NO"" because I locked myself out for other reasons, so I wanted to try it step by step this time.
Is the rule wrong or is logging without an enabled firewall the problem?
I have a pretty active phpBB forum. I was looking at my stats today and I notice some strange referrals. I checked it out and it and it is a bunch of spam sites. But with a link to one of my threads in the bottom (different per spam site).
- The spam sites have nothing related to my content - The link to the threads in my forum has no relation to each other. - Out of all the sites to put, why mine? My forum is not a very very popular forum.
I can post the link if it will better understand the above situation.
I just like to know why and what purpose does it serve? Also is it bad for me? As it is a spam site is linking to my forum. I dunno if search engines will think my site is a spam site as well.
Users of my web site, running MD-Pro, a PostNuke (php-based) clone, suddenly started to report difficulties downloading files from the download module. We assumed it was the download module and spent ages diagnosing it, uploading fresh versions etc. After some time someone had the bright idea to see if the files could be downloaded direct (using the URLs) to eliminate the web application.
We found: Some users can download all files from the server with no problems. Some users can download some files but not others.
So far as we can tell only MS Word files are affected, but not all.
The files on the server are not corrupted and there are no permissions issues
Disabling antivirus and firewalls locally makes no difference.
REloading fresh originals to the servedr does not help.
The host reports that no changes to the server have been made recently.
We have discovered that files in the web root are OK. The furhte down the directory structure they are, the less likely they are to download. Performance varies according to the browser in use. According to the host technical support, 'Firefox appears to be returning the data from its own cache. IE is only doing so for root. For the other places IE tries to download the file but stops receiving at exactly the same number of bytes from different locations.'
Using another server on a different host, the problem disappears so it must be due to the host setup in some way. Diagnosis is difficult because the host technical support can't reproduce it.
Platforms include Mac, Windows and Linux and browsers include IE7, Firefox and Safari.
One user with Mac isn't having problems, another with Mac is. All others reporting problems are on Windows. Users are at different locations, using different ISPs so it is unlikely there are common local problems.
The only common element I can see is that the ony files causing trouble are MS Word, though some of these do download OK
We have run out of ideas why this should be happening. How can some users have problems with the same files on the same server and others not?
What could have happened to cause this problem on a site that has been functioning correctly for several months on this server?
Feb 19 15:57:39 server proftpd[1363]: server.com (127.0.0.1[127.0.0.1]) - FTP session closed. Feb 19 16:06:02 server proftpd[1982]: server.com (127.0.0.1[127.0.0.1]) - FTP session opened. Feb 19 16:06:02 server proftpd[1982]: server.com (127.0.0.1[127.0.0.1]) - FTP session closed. Feb 19 16:14:24 server proftpd[2471]: server.com (127.0.0.1[127.0.0.1]) - FTP session opened. Feb 19 16:14:24 server proftpd[2471]: server.com (127.0.0.1[127.0.0.1]) - FTP session closed. Feb 19 16:22:46 server proftpd[3062]: server.com (127.0.0.1[127.0.0.1]) - FTP session opened. Feb 19 16:22:46 server proftpd[3062]: server.com (127.0.0.1[127.0.0.1]) - FTP session closed. Feb 19 16:31:09 server proftpd[3696]: server.com (127.0.0.1[127.0.0.1]) - FTP session opened. Feb 19 16:31:09 server proftpd[3696]: server.com (127.0.0.1[127.0.0.1]) - FTP session closed. Feb 19 16:39:31 server proftpd[4185]: server.com (127.0.0.1[127.0.0.1]) - FTP session opened. Feb 19 16:39:31 server proftpd[4185]: server.com (127.0.0.1[127.0.0.1]) - FTP session closed. Feb 19 16:47:53 server proftpd[4946]: server.com (127.0.0.1[127.0.0.1]) - FTP session opened. Feb 19 16:47:53 server proftpd[4946]: server.com (127.0.0.1[127.0.0.1]) - FTP session closed. Feb 19 16:56:16 server proftpd[5495]: server.com (127.0.0.1[127.0.0.1]) - FTP session opened. Feb 19 16:56:16 server proftpd[5495]: server.com (127.0.0.1[127.0.0.1]) - FTP session closed. Feb 19 17:04:38 server proftpd[6206]: server.com (127.0.0.1[127.0.0.1]) - FTP session opened. Feb 19 17:04:38 server proftpd[6206]: server.com (127.0.0.1[127.0.0.1]) - FTP session closed. Feb 19 17:13:00 server proftpd[6661]: server.com (127.0.0.1[127.0.0.1]) - FTP session opened. Feb 19 17:13:00 server proftpd[6661]: server.com (127.0.0.1[127.0.0.1]) - FTP session closed. Feb 19 17:21:23 server proftpd[7225]: server.com (127.0.0.1[127.0.0.1]) - FTP session opened. Feb 19 17:21:23 server proftpd[7225]: server.com (127.0.0.1[127.0.0.1]) - FTP session closed.
I see over a few hundred of these lines in /var/log/messages. The timestamp is exactly the same for every 2 lines (Proftp session- Opened and Closed). It's occuring every hour of the day. Is someone attacking the ftp daemon or something?
Ever since today my server is being very strange... The server load is always around 6-10 and sometimes goes up to 16! It never did something like this before and I can't find any weird Apache behavior either. Though when I shut down Apache the server load drops to below 1 so it must be Apache...
Kernel: Linux alpha.pixelhosting.net 2.6.9-42.0.10.ELsmp #1 SMP Tue Feb 27 10:11:19 EST 2007 i686 i686 i386 GNU/Linux
A client of mine contacted me to do some changes in his website that's being hosted with Godaddy (it was not me who did the previous works, I'm trying first time a site hosted with Godaddy).
So I tried to enter the site thro' my SmartFTP, but the moment I hit enter after filling up the address, login and password, the login and password fields turn greyish, and the address gets changed to 'domain name - anonymous'.
I get a small remote browser like I get in case of other wesites, but I get something empty in this case. I don't see any files of the website present in that browser.
