[Engine-devel] Gluster IPTable configuration

Sun Sep 2 14:21:47 UTC 2012

----- Original Message -----
> From: "Andrew Cathrow" <acathrow at redhat.com>
> To: "Alon Bar-Lev" <alonbl at redhat.com>
> Cc: "David Jaša" <djasa at redhat.com>, engine-devel at ovirt.org
> Sent: Friday, August 31, 2012 3:12:34 PM
> Subject: Re: [Engine-devel] Gluster IPTable configuration
> 
> 
> 
> ----- Original Message -----
> > From: "Alon Bar-Lev" <alonbl at redhat.com>
> > To: "David Jaša" <djasa at redhat.com>
> > Cc: engine-devel at ovirt.org
> > Sent: Friday, August 31, 2012 5:09:47 AM
> > Subject: Re: [Engine-devel] Gluster IPTable configuration
> > 
> > 
> > 
> > ----- Original Message -----
> > > From: "David Jaša" <djasa at redhat.com>
> > > To: engine-devel at ovirt.org
> > > Sent: Friday, August 31, 2012 11:57:11 AM
> > > Subject: Re: [Engine-devel] Gluster IPTable configuration
> > > 
> > > Alon Bar-Lev píše v Čt 30. 08. 2012 v 14:40 -0400:
> > > > 
> > > > ----- Original Message -----
> > > > > From: "Andrew Cathrow" <acathrow at redhat.com>
> > > > > To: "Alon Bar-Lev" <alonbl at redhat.com>
> > > > > Cc: "Shireesh Anjal" <sanjal at redhat.com>,
> > > > > engine-devel at ovirt.org,
> > > > > "Selvasundaram" <sesubram at redhat.com>
> > > > > Sent: Thursday, August 30, 2012 9:37:59 PM
> > > > > Subject: Re: [Engine-devel] Gluster IPTable configuration
> > > > > 
> > > > > 
> > > > > 
> > > > > ----- Original Message -----
> > > > > > From: "Alon Bar-Lev" <alonbl at redhat.com>
> > > > > > To: "Selvasundaram" <sesubram at redhat.com>
> > > > > > Cc: "Shireesh Anjal" <sanjal at redhat.com>,
> > > > > > engine-devel at ovirt.org
> > > > > > Sent: Thursday, August 30, 2012 2:35:16 PM
> > > > > > Subject: Re: [Engine-devel] Gluster IPTable configuration
> > > > > > 
> > > > > > 
> > > > > > 
> > > > > > ----- Original Message -----
> > > > > > > From: "Selvasundaram" <sesubram at redhat.com>
> > > > > > > To: engine-devel at ovirt.org
> > > > > > > Cc: "Shireesh Anjal" <sanjal at redhat.com>
> > > > > > > Sent: Thursday, August 30, 2012 4:30:16 PM
> > > > > > > Subject: [Engine-devel] Gluster IPTable configuration
> > > > > > > 
> > > > > > > 
> > > > > > > Hi,
> > > > > > > 
> > > > > > > I want to add gluster specific IPTable configuration in
> > > > > > > addition
> > > > > > > to
> > > > > > > the ovirt IPTable configuration (if it is gluster node).
> > > > > > > 
> > > > > > > There are two approaches,
> > > > > > > 1. Having one more gluster specific IP table config in db
> > > > > > > and
> > > > > > > merge
> > > > > > > with ovirt IPTable config (merging NOT appending)
> > > > > > > [I have the patch engine: Gluster specific firewall
> > > > > > > configurations
> > > > > > > #7244]
> > > > > > > 2. Having two different IP Table config (ovirt and
> > > > > > > ovirt+gluster)
> > > > > > > and
> > > > > > > use either one.
> > > > > > > 
> > > > > > > Please provide your suggestions or improvements on this.
> > > > > > > 
> > > > > > 
> > > > > > Hello all,
> > > > > > 
> > > > > > The mentioned patch[1], adds hard coded gluster code into
> > > > > > the
> > > > > > bootstrap code, manipulate the firewall configuration to be
> > > > > > gluster
> > > > > > specific. It hardcoded search for "reject", insert before
> > > > > > some
> > > > > > other
> > > > > > rules.
> > > > > > 
> > > > > > I believe this hardcode approach is obsolete now that we
> > > > > > have
> > > > > > proper
> > > > > > tools for templates.
> > > > > > 
> > > > > > A more robust solution would be defining generic profiles,
> > > > > > each
> > > > > > profile as a template, each template can refer to different
> > > > > > profiles, and assign profile to a node.
> > > > > > 
> > > > > > This way the implementation is not gluster [or any]
> > > > > > specific
> > > > > > and
> > > > > > can
> > > > > > be reused for more setups, code is cleaner.
> > > > > 
> > > > > 
> > > > > or create custom chains ?
> > > > 
> > > > Can you please elaborate what is custom chains?
> > > > Thanks!
> > > 
> > > iptables -N my_new_chain
> > > iptables -A my_new_chain <rule_1>
> > > iptables -A my_new_chain ...
> > > iptables -A my_new_chain <rule_n>
> > > 
> > > # if this <rule> is matched, packet goes through rules in
> > > my_new_chain
> > > iptables -A INPUT <rule> -j my_new_chain
> > > 
> > 
> > Hello,
> > 
> > How does this solve the original issue?
> 
> It makes it easier for customers who are adding their own IPTables
> configuration - when we do rhev-h plugins it'll make things easier.
> This way we don't wipe out the original rules we just add our own
> chain.
> 
> 

Current default (virt) iptables configuration is kept in the engine's DB.
So...

1. We can update the DB value to include Gluster configuration.
or
2. Have an additional DB value for virt+Gluster.

So (1) will work, but when gluster is not relevant we have redundant ports.
But my only reservation from (2) is that it's not scalable for additional
services.

Either way, I agree with Alon that melding the configurations at runtime
is unexpected, and shouldn't be used.

After some additional thought... we can have (1), and make all gluster
entries commented out (let's say with #GLSTR) by default, so it'll be
enabled only when gluster is being used. This will allow introducing future
configurations as well without the need for additional DB values.

Sample snip of DB value:
==================================================================================
:INPUT ACCEPT [0:0]
:FORWARD ACCEPT [0:0]
:OUTPUT ACCEPT [0:0]
-A INPUT -m state --state ESTABLISHED,RELATED -j ACCEPT
-A INPUT -p icmp -j ACCEPT
-A INPUT -i lo -j ACCEPT
# vdsm
-A INPUT -p tcp --dport 54321 -j ACCEPT
# libvirt tls
-A INPUT -p tcp --dport 16514 -j ACCEPT
# SSH
-A INPUT -p tcp --dport 22 -j ACCEPT
# guest consoles
-A INPUT -p tcp -m multiport --dports 5634:6166 -j ACCEPT
# gluster p1
#GLSTR-A INPUT -p tcp --dport p1 -j ACCEPT
# gluster p2-p6
#GLSTR-A INPUT -p tcp --dport p2:p8 -j ACCEPT
# Reject any other input traffic
-A INPUT -j REJECT --reject-with icmp-host-prohibited
-A FORWARD -m physdev ! --physdev-is-bridged -j REJECT --reject-with icmp-host-prohibited
COMMIT
==================================================================================

So as you can see, this is something we can handle in run-time. Also it's easy
to test, so predictable, and finally- human beings can look and understand what
it means.

> > The need to provide different rules to different hosts by software
> > installed on destination?
> > 
> > Standard host needs iptables X.
> > Gluster host needs iptables X+Y.
> > XXX host needs iptables X+Z.
> > Maintainer of Gluster knows what Y is.
> > Maintainer of XXX knows what Z is.
> > 
> > If we merge all to one entry product comes with default X.
> > User override X to A.
> > New version of product comes with default Y.
> > Upgrade options:
> > 1. System continues to use A.
> > 2. Some AI to upgrade and create A'.
> > 3. Revert to Y, dropping user's customization.
> > 
> > Or we can maintain one large table with complete configuration and
> > conditionals.
> > 
> > Alon.