Evgeny's questions

I've tried to install mod_passenger on a 64 bit CentOS5.3 machine and the module compiles into a 32 bit .so file. Followed standard recipe: gem install passenger; passenger-install-apache2-module.

• apache 2.2.3
• Rails 2.3.5
• gem 1.35

The passenger-install-apache2-module command emits a script (while running it) with lot's of flags like -m32 and -march=i386. I've captured the script, replaced -m32 with -m64 and -march=i386 with -march=x86_64, re-ran it (with some extra include path and APXS2 variables) but the output object is still in 32 bit format.

Could someone share insight on how to force the 64-bitness upon the passenger?

edit: later I found out that any module I would try to compile would be built into a 32 bit version, so it's not specific to passenger

Thanks!

would anyone give insight on this datetime issue:

My web-application (on centos/apache/wsgi/django) shows different current time (datetime.datetime.now()) from call to call - I'm not asking why the time keeps going ahead :)

I get the following in my debugging log: (please notice that first and last entries are made by the same process, but the latter one appears to be created five hours before the first one, which is nonsense)

2009-10-28 22:00:37,724 DEBUG pid=16436 time=2009-10-28 22:00:37.724682 
2009-10-28 22:00:40,194 DEBUG pid=16407 time=2009-10-28 22:00:40.194602 
2009-10-28 22:00:42,742 DEBUG pid=16409 time=2009-10-28 22:00:42.742598 
2009-10-28 22:00:44,643 DEBUG pid=16408 time=2009-10-28 22:00:44.643639 
2009-10-28 22:00:46,747 DEBUG pid=16436 time=2009-10-28 22:00:46.747777 
2009-10-28 22:00:48,973 DEBUG pid=16411 time=2009-10-28 22:00:48.973176 
2009-10-28 22:00:51,261 DEBUG pid=16410 time=2009-10-28 22:00:51.261172 
2009-10-28 17:13:16,867 DEBUG pid=16436 time=2009-10-28 17:13:16.867307

Thanks!

tried to install squid with command

yum install squid

and it consistently hangs on

Running Transaction
 Installing: squid                        #####################     [1/1]

Had no problems with hundreds of other packages. Does anyone have an idea on what could be the problem?

edit: tried both: yum clean all + changing mirror, and rebuild rpm, doesn't help.

Thanks.

I would like to copy stuff in bulk (reimage disk using dd) with netcat from host A to B via ssh encrypted channel on Linux.

What commands should I type on both ends?

I need to migrate my CentOS server onto another with an identical system distribution. I don't know yet how different is hardware the besides amount of memory and GHz (posted separate question about that)

I've been compiling lot's of libraries by hand on my original server, installing them to standard location and did not keep a record :). (there is no one to fire me for this - I only lose time and money here and hopefully - learn something)

Migration will also involve moving user stuff, setting permissions, file ACLs. (here is question on moving user accounts, but there's nothing about facl)

Is there anything in the world that can help me re-create the beast on the new host?

Are there any tools that allow such migrations in a more orderly fashion? (preventive or more active tools)

Obviosuly I'm no sysadmin - but I hope to learn more. Thanks.

There was a similar question about migrating windows server - in my case it's linux.

-i.e - how to get a full list of hardware components in command line (on a machine with no window system)

Thank you.

Want to make pages load in a blink.

My front page takes 1.1 s to load with ping 74 ms. A similar (ping 72 ms) English wikipedia page (actually 40% larger document on a heavily used host) loads in 0.4 s. My site runs Mediawiki software under APC on a dedicated host with 1 GHz Athlon, there barely is any traffic and the machine has some free core memory.

Is there a way to speed up the load, or anything I could look at? I've tried squid and there was no improvement at all.

The test I used was time wget url

Thanks.

P.S. I've asked similar question, but I think it was not stated clearly enough.

Basically I'd like to know if it's possible to speed up my server or not by using caching like squid, checking for possible errors in configuration, optimizing various server software parameters, etc.

Is it possible to predict time to serve say 1MB of data if I know ping time and clock speed of my server, and other hardware parameters.

for example in my case I have 1GHz AMD with 75ms ping and I get 16 kB page served from core memory in 0.6s - can I make it any faster? - just plain time to display page at no load.

Do any benchmarks/tools exist for this purpose?

edit I've tried adding squid and did not get any performance gain at all.

Thanks.

... related to a previous question about trouble running postfix on Centos 5.2(Final)

Looks like postfix relies on aliases.db which is built by newaliases.postfix command

When I run newaliases.postfix it just hangs indefinitely. I have found some tips, but they weren't helpful (I've changed /etc/hosts and /etc/sysconfig/network accordingly to no effect)

Running strace newaliases.postfix shows this

...
close(4)                                = 0
munmap(0x2b9bb210c000, 4096)            = 0
read(0,  <unfinished ...>

Apparently it's trying to read from standard input but never gets any input.

Am I supposed to feed something to newaliases?

When I try newaliases.postfix < /etc/aliases I get

newaliases.postfix: fatal: Recipient addresses must be specified on \
the command line or via the -t option

I've just compiled and installed postfix and set domain names in main.cf.

Unfortunately postfix does not start. Nothing new shows up in the list of processes when I issue /usr/sbin/postfix start

echo $? gives 0

Are there any log files or other clues I could use to figure out what goes wrong? Thanks.

Which processes should I look for in the ps output?

Thanks.

setting up a simple mailserver to forward all messages sent to one address to gmail.

So far I have a file /var/qmail/mailnames/myserver.com/username/.qmail with a single line:

&[email protected]

I also have file /var/qmail/users/assign with:

=username:localuser:2525:48:/var/qmail/mailnames/servername.com/username:::

test email bounces with error "Sorry, no mailbox here by that name."

Thanks!

my mod_wsgi/django app has ~2s response time with or without memcached in-memory caching.

response is the same 2s with or without DNS resolution (i.e. with direct IP request)

for comparison, a simulaneously running mediawiki php process running with APC has sub-second response.

the machine is not out of memory and I can tell that memcached uses up some space.

I've seen the same django app run twice faster on some other machine (of which I unfortunately know little details, but it seems to be running on webfactional) my machine is dedicated host.

here is my apache directive to run the app

<VirtualHost ...:80>
     ServerAdmin ...
     DocumentRoot ...
     ServerName ..
     WSGIScriptAlias / /../wsgi-script.wsgi
     CustomLog ... common
     ErrorLog /.../error_log
     DirectoryIndex index.html
</VirtualHost>

EDIT ok found the source of the problem - wsgi was serving static files and that's what drained the performance. Some Alias-type directives for static content fix the issue.