theist's questions

I'm having a weird case of syslog-ng freezing on some reloads and under some conditions.

syslog-ng version is ubuntu 20.04's default 3.25.1. It has to be something related to log_msg_size as this started to happen when it was reconfigured from log_msg_size(16384) to log_msg_size(65536) which, in fact, is its default value.

Once it was reconfigured the daemon was left in state Reloading configuration according to systemd, and stopped login at all. Also, it was not responding to TERM signal having to forcefully kill it. In this state any attempt to reload config got an error warning about an already existing config reloading.

If this was not oddly enough, killing the daemon, starting it with log_msg_size(16384) and then stopping it normally and reconfiguring it again to log_msg_size(65536) makes the reload work correctly until something happens and the reload start to hang the process in the described state.

Any clues of what I could be missing? I've tried to remove the option completely and let it get its own defaults, but it seems that still happens the same.

Is there a way to tell the kernel to keep a certain amount of free RAM and do not use it for caches?

I hear you thinking, "Why you would want that?! cached RAM is free RAM when someone wants it!"

This is a special case. It seems that a nginx plugin for an ip location service check the amount of free ram before reload and it refuses to restart / reload if there's not enough strictly free RAM to do it. And I have a couple of servers in this situation... all RAM cached and a nginx that refuses to restart itself. So I have to manually (or "cronly") do a drop_caches to have enough strict free RAM for the nginx restart.

I'd like to have a way to specify a max amount of cached ram or a minimum strict free RAM to keep available to get rid of the hacky drop_caches method

I'm wondering if is it possible to specify another main.cf for the configuration of smptd service in another port paired with normal config in standard port.

I know that is it possible to override main.cf settings in master.cf using the -o <key>=<value> parameter for smtpd service definitions but I'm asking for a way to directly tell to smtpd to read another main configuration file.

Can it be done?

It's possible to force the re-creation of a EC2 or RDS instance using cloudformation stacks?

My stack goes stuck in a point where simply destroying and creating the resource will fix it, instead of that I had to delete entire stack to continue work.

edit:

This issue hit me twice. First I created an AWS::RDS::Instance with some defaults and then tried to downgrade it to "EngineVersion" : "5.5". Changing this its suposed to happen with some interruption, but mysql instances cannot be downgraded from 5.6 to 5.5 so the stack was left in UPDATE_FAILED state and I cannot be able to recreate RDS without a nasty trick.

The other occurrence was that i have several "AWS::EC2::Instance" which downloads and executes an script from it's "UserData" obviously if Y change the downloaded script I must recrete the instance, and theres no way to do so. Once again I use the same nasty trick to get the machine recreated.

The nasty trick:

Instead of using an autoscaling group of one machine, I solved both problems changing the availability zone in the properties... but left me with a bad taste

I'm using qemu/kvm whith bridged networking. In the host machine there are several "vnetX" network interfaces without IP. I'm looking for a way to know which vnetX belong to a virtual machine.

I tried to match the MAC Address values on these interfaces with the MACs on the virtual machines (or the XML which defines them), but doesn't match.

There's brctl show which shows the vnet interfaces that belongs to a bridge, but this is not useful info.

Is there a way to know that relation? Thx!!

I'm used to use /etc/hosts to fake "example.com" or other domains to point to my local web servers to make tests.

I'm trying to do some filtering tests on a multi-domain postfix and I need to fake several domain names to check how it responds without touching the real thing.

Is possible to do so, in a easy way, to make configuration test with a local mail server?

If it is not as easy tell me how to do it also :)

Hi I'm trying to incorporate a software to our intranet services (BackupPc)

This Software uses the environment variable REMOTE_USER to get the username. Placed under an Apache 2.2 server with Client certificate Authentication system and FakeBasicAuth enabled, REMOTE_USER gets the same value as SSL_CLIENT_S_DN, which have all the distinguished name string of the certificate.

I'm trying to set REMOTE_USER the contents of SSL_CLIENT_S_DN_CN which has a content which matches with my usernames and easyer and shorter to read.

This environment variable passed to CGI can not be changed by SetEnv and SetEnfIf. There's also a mod_ssl directive, SSLUserName but the Apache 2.2 official documentation says that it won't work with FakeBasicAuth enabled.

So I'm wondering if I'm missing something or there's a clever workarround on this issue.

I can't drop FakeBasicAuth because that's our authmodel on the rest of the intranet, And I dont want to patch the code of the application (I did it to get it working but I want to drop the patched code in favor of the official one)

I need to restrict certain user from a BasicAuth configured system to come only from certain IP address but allow the rest to come from anywere.

I know that I can restrict access using Deny and Allow directives and know also how to restrict access to users using Require. But I need a mix of both and I don't have a clue.

Is there a way to get a command line prompt to a hosted virtual machine in vmware server like xm console <name> does in Xen?

I know the graphic UI to conect to vmware-servers but I'm under low latency ssh only access and I need something more "text friendly"

I've got a server with several scponly accounts and I want to provide http access under Basic Auth.

The first part of the problem is solved, using mod_rewrite to chop and select the right directory for each user. But the problem arise when I try to stablish a different auth user file for each user. Simply I can't find how to do it.

Here's a sample config.

<VirtualHost an-ip-number-and:a-port>
    ServerName *.example.com
    RewriteEngine On
    UseCanonicalName Off

    RewriteCond %{HTTP_HOST} ^(.*).example.com
    RewriteCond /server/scponly/%1/incoming -d
    RewriteRule ^(.+)   %{SERVER_NAME}$1 [C]
    RewriteRule ^([^.]+)\.example\.com/(.*) /server/scponly/$1/incoming/$2 [L]

    RewriteRule ^(.+) http://FAIL.example.com

    <Directory /server/scponly/*/incoming>
        AuthType Basic
        AuthName "SFTP-HTTP Area"
        AuthUserFile  ???????   # <-- this!
        require valid-user
    </Directory>
</VirtualHost> 

Notes:

• I can't drop mod_rewrite beacuse actual config is mixed with other servername / directory-check / rewrites for other purposes.
• I can't move auth details to .htaccess files because users could erase them. But I'm interested if this approach would be a partial solution.
• I'm also interested on other approaches
• I need differenciate auth files for each directory, having only one file will grant access to all directories to each user