Re: [vox-tech] Risks of upgrading past CentOS 6 supported PHP 5.4?

2016-06-03 Thread Rick Moen
Quoting Dr. Larry Ozeran ([email protected]):

> Thanks Rick.
> 
> This is why I have such great respect for the members of this list.
> You have such valuable experiences that you are willing to share. I
> regret that I have had the experience of server issues occurring at
> bad times (right after talking about our product at a trade event),
> but thus far none have been PHP or MySQL related, so I very much
> appreciate the insights of others.

Looking over my notes about tightening PHP security (direct link
http://linuxmafia.com/faq/Security/php.html), I see something I might
consider writing more about  (in that piece):  I talk about several
prototype files Debian provided for /etc/php5/apache2/php.ini ,
specifically:

/usr/share/doc/php5-common/examples/php.ini-paranoid
/usr/share/doc/php5-common/examples/php.ini-recommended

The thing is, the distribution-provided  /etc/php5/apache2/php.ini 
file stated in header comment lines... well, let me just quote it:

;;;
; About this file ;
;;;
; PHP comes packaged with two INI files. One that is recommended to be used
; in production environments and one that is recommended to be used in
; development environments.

; php.ini-production contains settings which hold security, performance and
; best practices at its core. But please be aware, these settings may break
; compatibility with older or less security conscience applications. We
; recommending using the production ini in production and testing environments.

; php.ini-development is very similar to its production variant, except it's
; much more verbose when it comes to errors. We recommending using the
; development version only in development environments as errors shown to
; application users can inadvertently leak otherwise secure information.


So, thing is, I could swear that the _default_ php.ini Debian installed
to /etc/php5/apache2/php.ini was the _development_ variant -- because
when I got around to looking through it and locking it down, I was a bit
shocked at the high default security exposure.  It was exactly as I was
working through that file, commenting out dangerous and unneeded
optional interpreter features, that I wrote my modest hardening guide
(more as notes for my own benefit than with any intent to publish).

That experience was one of my first signs that something was not right
with PHP and its security posture.  It might be my imagination, but 
I think part of what's happening is that only lip service is normally
given to enabling only the language functionality actually necessary; 
that many developers of PHP developed applications blithely expect 
deploying sites to have the kitchen-sink set of language functions
available and their code mysteriously fails if you don't.  So, deploying
sites are passively led into bad habits and bad defaults.

And, for example, picture a world where the developers gravitate towards
php.ini-development and sysadmins towards php.ini-production.  What's
going to happen as a natural tendency?  You know developers; they'll
happily use the features in front of them, which then will ratchet up
the need for those in deployment.  Not good.  It's a bad sign when the
development default environment includes bad security and the attitude
is 'be unsafe only while developing code and we'll fix it all up later'.

A brief clarification:  The incursion that happened just as I got on an
ocean cruise this past February happened _despite_ all the security
tightening, on account of yet another grave flaw in the core mod_php
code.  (I cannot tell you which flaw because I elected to just disable
the thing rather than work further on it.)

___
vox-tech mailing list
[email protected]
http://lists.lugod.org/mailman/listinfo/vox-tech


Re: [vox-tech] Risks of upgrading past CentOS 6 supported PHP 5.4?

2016-06-02 Thread Bill Broadley



Somehow with a painfully thin straw of ssh bandwidth and only one hour
of high-latency, low-reliability Internet access each evening, I was
able to kludge together a lockout of the kiddies within a couple of
days and before they were able to compile an exploit kit.  When I
reached Sydney, one of the first things I did from my hotel room was rip
out the last bits of public-facing PHP exposure so I'd never have to
worry about that again.


BTW, anyone in this situation should immediately install mosh .  It's 
pretty amazingly better, can even fall over between multiple network 
connections.


Or install it now just in case you have to tether to a marginal 
cellular/wan connection.


Blurb from https://mosh.mit.edu/:

  Remote terminal application that allows roaming, supports
  intermittent connectivity, and provides intelligent local echo and
  line editing of user keystrokes.

  Mosh is a replacement for SSH. It's more robust and responsive,
  especially over Wi-Fi, cellular, and long-distance links.

  Mosh is free software, available for GNU/Linux, FreeBSD, Solaris, Mac
  OS X, and Android.
___
vox-tech mailing list
[email protected]
http://lists.lugod.org/mailman/listinfo/vox-tech


Re: [vox-tech] Risks of upgrading past CentOS 6 supported PHP 5.4?

2016-06-02 Thread Dr. Larry Ozeran

Thanks Rick.

This is why I have such great respect for the members of this list. You 
have such valuable experiences that you are willing to share. I regret 
that I have had the experience of server issues occurring at bad times 
(right after talking about our product at a trade event), but thus far 
none have been PHP or MySQL related, so I very much appreciate the 
insights of others.


Thanks again,

Dr. Larry Ozeran
President, Clinical Informatics, Inc.
(530) 671-9244

On 6/2/2016 18:46, Rick Moen wrote:

Quoting Dr. Larry Ozeran ([email protected]):


Rick, thanks again for your insights.

You are most welcome.
  

You are, of course, correct that we would not redesign our software
without a significant and deep assessment of benefits and costs
(money, time, resources, etc.). Most of the PHP, MySQL, and related
code has been developed in house. I probably coded 10-15% myself.
The intent of my comment was simply to indicate that we do not
blindly accept that there is no better option than what we are
doing. If there are strong arguments to support considering making a
switch, I would not exclude that possibility without reviewing the
pros and cons simply because we have a large legacy investment. I
consider your response (below) to fall into the 'cons' (to
switching) category and will definitely compare your PHP security
recommendations against what we currently are and are not doing.

I am very glad to be of help -- and certainly was trying to be at pains
to avoid advising anyone to merely redesign, especially without
knowledge of the particulars.

My own disaffection with PHP was markedly increased when I boarded a
cruise ship with my wife from San Francisco to Sydney, and right on the
day of my departure my logcheck reports started indicating a serious
attempt to break security on my server via (what turned out to be)
mod_php -- exactly at a time when I had just boarded an ocean vessel
with only satellite Internet at very high prices.

Somehow with a painfully thin straw of ssh bandwidth and only one hour
of high-latency, low-reliability Internet access each evening, I was
able to kludge together a lockout of the kiddies within a couple of
days and before they were able to compile an exploit kit.  When I
reached Sydney, one of the first things I did from my hotel room was rip
out the last bits of public-facing PHP exposure so I'd never have to
worry about that again.

My _own_ view is that PHP is entirely too much like the scenario
Marcus Ranum described in his rather caustic 'What Sun Tsu Would Say'
essay, i.e., as Ranum phrases it, 'If patching hasn't been working, why
are we still doing it?'  I stopped needing to apply the PHP patch du
jour by no longer exposing it to public networks.

But whatever works for you is of course great.

___
vox-tech mailing list
[email protected]
http://lists.lugod.org/mailman/listinfo/vox-tech


___
vox-tech mailing list
[email protected]
http://lists.lugod.org/mailman/listinfo/vox-tech


Re: [vox-tech] Risks of upgrading past CentOS 6 supported PHP 5.4?

2016-06-02 Thread Rick Moen
Quoting Dr. Larry Ozeran ([email protected]):

> Rick, thanks again for your insights.

You are most welcome.
 
> You are, of course, correct that we would not redesign our software
> without a significant and deep assessment of benefits and costs
> (money, time, resources, etc.). Most of the PHP, MySQL, and related
> code has been developed in house. I probably coded 10-15% myself.
> The intent of my comment was simply to indicate that we do not
> blindly accept that there is no better option than what we are
> doing. If there are strong arguments to support considering making a
> switch, I would not exclude that possibility without reviewing the
> pros and cons simply because we have a large legacy investment. I
> consider your response (below) to fall into the 'cons' (to
> switching) category and will definitely compare your PHP security
> recommendations against what we currently are and are not doing.

I am very glad to be of help -- and certainly was trying to be at pains
to avoid advising anyone to merely redesign, especially without
knowledge of the particulars.

My own disaffection with PHP was markedly increased when I boarded a
cruise ship with my wife from San Francisco to Sydney, and right on the
day of my departure my logcheck reports started indicating a serious 
attempt to break security on my server via (what turned out to be)
mod_php -- exactly at a time when I had just boarded an ocean vessel
with only satellite Internet at very high prices.

Somehow with a painfully thin straw of ssh bandwidth and only one hour
of high-latency, low-reliability Internet access each evening, I was
able to kludge together a lockout of the kiddies within a couple of
days and before they were able to compile an exploit kit.  When I
reached Sydney, one of the first things I did from my hotel room was rip
out the last bits of public-facing PHP exposure so I'd never have to
worry about that again.

My _own_ view is that PHP is entirely too much like the scenario 
Marcus Ranum described in his rather caustic 'What Sun Tsu Would Say'
essay, i.e., as Ranum phrases it, 'If patching hasn't been working, why
are we still doing it?'  I stopped needing to apply the PHP patch du
jour by no longer exposing it to public networks.

But whatever works for you is of course great.

___
vox-tech mailing list
[email protected]
http://lists.lugod.org/mailman/listinfo/vox-tech


Re: [vox-tech] Risks of upgrading past CentOS 6 supported PHP 5.4?

2016-06-02 Thread Dr. Larry Ozeran
Thanks Alex. I have heard about Docker but no one on our team has 
experience with it. I will look into it.


Dr. Larry Ozeran
President, Clinical Informatics, Inc.
(530) 671-9244

On 6/2/2016 07:35, Alex Mandel wrote:

This might be a a good use case for a Docker container, which is
essentially all the components you need but not messing with your system
libraries and dependencies.

If you're open to other options, python+ django/pyramid/ or flask, +
postgres/,mysql, sqlite. Or Ruby + Rails + Database.

Those are 2 very common alternatives to PHP based stuff.

Enjoy,
Alex

On 06/02/2016 05:39 AM, Dr. Larry Ozeran wrote:

Thanks Rick. Good information is always appreciated.

Since we are serving data that can change every few minutes, we can't
move to static pages. Since we are providing that data to users from
multiple originating sources, we pretty much have to be internet-facing.
We have put security procedures in place, but I know that security is
more an ongoing process than an endpoint and there is always more that
will need to be done. If there is a better way to meet the needs of
users other than MySQL+PHP, I am always open to new ideas.

Thanks,

Dr. Larry Ozeran
President, Clinical Informatics, Inc.
(530) 671-9244

On 6/2/2016 00:41, Rick Moen wrote:

Quoting Bill Broadley ([email protected]):


Does anyone know any downsides to using the webtatic PHP packages on
CentOS 6?

I've seen many machines with ugly configurations related to cpanel,
custom php installs (sometimes more than one), and fragile very hard
to reproduce apache configurations.

Although I guess I shouldn't complain, they get hacked and I get
consulting.

(Sadly, this won't answer Dr. Ozeran's question, either:)

I've lately come to the conclusion, from many years as a Linux server
admin, that PHP is tolerable on a Unix machine _provided_ it isn't ever
exposed to public networks, because the ongoing security nightmare is
otherwise not justifiable.  I mean, yes if management is paying you to
do it and the money's good, but if you're the boss, say 'Hell no.'

So, e.g., every Web page on my linuxmafia.com server that used to be
dymanically assembled by the PHP interpreter at Apache page-load time
are (more recently) instead built on-disk in advance using automake or a
cron script.  Fortunately, none of those pages needed to _actually_ be
dynamic; it was just coder laziness that chose that implementation.

For example, the coder who helped me convert BALE
(http://linuxmafia.com/bale/) from its original mid-1990s static HTML
incarnation to PHP + MySQL set it up so every page load assembles the
page anew, from several PHP fragments plus the results of a MySQL query
(furnishing the events rows).  When I realised the underlying reality of
this being static data changing only once on the 1st of each month, I
converted it into a static HTML page generated by a cron job in
/etc/cron.monthly/ , and then Apache serves up just that static file.

Whole huge categories of security threat have completely away for good,
when I ditched runtime PHP.

If I had any Web applications that actually relied on the PHP
interpreter at load time, I'd try really hard to ditch them.  It really
is IMO that bad.

And I say that because, so to speak, Ranum is my guru:
http://www.ranum.com/security/computer_security/editorials/master-tzu/

That having been said:  Dr. Ozeran, I know of nothing against the
Webtatic repo's PHP packages.  It seems like a competent external repo
for CentOS/RHEL, though I have no relevant experience.  Hope that helps!


___
vox-tech mailing list
[email protected]
http://lists.lugod.org/mailman/listinfo/vox-tech

___
vox-tech mailing list
[email protected]
http://lists.lugod.org/mailman/listinfo/vox-tech

___
vox-tech mailing list
[email protected]
http://lists.lugod.org/mailman/listinfo/vox-tech


___
vox-tech mailing list
[email protected]
http://lists.lugod.org/mailman/listinfo/vox-tech


Re: [vox-tech] Risks of upgrading past CentOS 6 supported PHP 5.4?

2016-06-02 Thread Dr. Larry Ozeran

Rick, thanks again for your insights.

You are, of course, correct that we would not redesign our software 
without a significant and deep assessment of benefits and costs (money, 
time, resources, etc.). Most of the PHP, MySQL, and related code has 
been developed in house. I probably coded 10-15% myself. The intent of 
my comment was simply to indicate that we do not blindly accept that 
there is no better option than what we are doing. If there are strong 
arguments to support considering making a switch, I would not exclude 
that possibility without reviewing the pros and cons simply because we 
have a large legacy investment. I consider your response (below) to fall 
into the 'cons' (to switching) category and will definitely compare your 
PHP security recommendations against what we currently are and are not 
doing.


Thanks,

Dr. Larry Ozeran
President, Clinical Informatics, Inc.
(530) 671-9244

On 6/2/2016 07:02, Rick Moen wrote:

Quoting Dr. Larry Ozeran ([email protected]):


Since we are serving data that can change every few minutes, we
can't move to static pages. Since we are providing that data to
users from multiple originating sources, we pretty much have to be
internet-facing. We have put security procedures in place, but I
know that security is more an ongoing process than an endpoint and
there is always more that will need to be done. If there is a better
way to meet the needs of users other than MySQL+PHP, I am always
open to new ideas.

Meaning no criticism, I notice in looking upthread
(http://lists.lugod.org/pipermail/vox-tech/2016-May/017013.html) that
you mention only that your use-case involves PHP-served pages, but not
what drives that particular choice of software.

Sometimes, a local site uses PHP because it runs developed software
resting on the PHP interpreter, e.g. Wordpress, MediaWiki, etc.
Other times, that choice resulted from 'Data for each page must be
pulled on a per-visit basis from MySQL, therefore some HTTP-invoked
process must do a SQL query and assemble page contents and we happened
to use PHP to do that because our Web guy knew how to do that.'  And
I'm sure there are other scenarios -- but dynamic is not synomyous with
PHP in any event.

Irrespective of how you arrived at that choice, obviously you would not
lightly decide to rearchitect.

A number of guides to tigthening PHP security to reduce risk exist and
may be useful.  My own modest effort, last updated when PHP5 was new, is
here:  'PHP Security' on http://linuxmafia.com/kb/Web/ .

___
vox-tech mailing list
[email protected]
http://lists.lugod.org/mailman/listinfo/vox-tech


___
vox-tech mailing list
[email protected]
http://lists.lugod.org/mailman/listinfo/vox-tech


Re: [vox-tech] Risks of upgrading past CentOS 6 supported PHP 5.4?

2016-06-02 Thread Alex Mandel
This might be a a good use case for a Docker container, which is
essentially all the components you need but not messing with your system
libraries and dependencies.

If you're open to other options, python+ django/pyramid/ or flask, +
postgres/,mysql, sqlite. Or Ruby + Rails + Database.

Those are 2 very common alternatives to PHP based stuff.

Enjoy,
Alex

On 06/02/2016 05:39 AM, Dr. Larry Ozeran wrote:
> Thanks Rick. Good information is always appreciated.
> 
> Since we are serving data that can change every few minutes, we can't
> move to static pages. Since we are providing that data to users from
> multiple originating sources, we pretty much have to be internet-facing.
> We have put security procedures in place, but I know that security is
> more an ongoing process than an endpoint and there is always more that
> will need to be done. If there is a better way to meet the needs of
> users other than MySQL+PHP, I am always open to new ideas.
> 
> Thanks,
> 
> Dr. Larry Ozeran
> President, Clinical Informatics, Inc.
> (530) 671-9244
> 
> On 6/2/2016 00:41, Rick Moen wrote:
>> Quoting Bill Broadley ([email protected]):
>>
 Does anyone know any downsides to using the webtatic PHP packages on
 CentOS 6?
>>> I've seen many machines with ugly configurations related to cpanel,
>>> custom php installs (sometimes more than one), and fragile very hard
>>> to reproduce apache configurations.
>>>
>>> Although I guess I shouldn't complain, they get hacked and I get
>>> consulting.
>> (Sadly, this won't answer Dr. Ozeran's question, either:)
>>
>> I've lately come to the conclusion, from many years as a Linux server
>> admin, that PHP is tolerable on a Unix machine _provided_ it isn't ever
>> exposed to public networks, because the ongoing security nightmare is
>> otherwise not justifiable.  I mean, yes if management is paying you to
>> do it and the money's good, but if you're the boss, say 'Hell no.'
>>
>> So, e.g., every Web page on my linuxmafia.com server that used to be
>> dymanically assembled by the PHP interpreter at Apache page-load time
>> are (more recently) instead built on-disk in advance using automake or a
>> cron script.  Fortunately, none of those pages needed to _actually_ be
>> dynamic; it was just coder laziness that chose that implementation.
>>
>> For example, the coder who helped me convert BALE
>> (http://linuxmafia.com/bale/) from its original mid-1990s static HTML
>> incarnation to PHP + MySQL set it up so every page load assembles the
>> page anew, from several PHP fragments plus the results of a MySQL query
>> (furnishing the events rows).  When I realised the underlying reality of
>> this being static data changing only once on the 1st of each month, I
>> converted it into a static HTML page generated by a cron job in
>> /etc/cron.monthly/ , and then Apache serves up just that static file.
>>
>> Whole huge categories of security threat have completely away for good,
>> when I ditched runtime PHP.
>>
>> If I had any Web applications that actually relied on the PHP
>> interpreter at load time, I'd try really hard to ditch them.  It really
>> is IMO that bad.
>>
>> And I say that because, so to speak, Ranum is my guru:
>> http://www.ranum.com/security/computer_security/editorials/master-tzu/
>>
>> That having been said:  Dr. Ozeran, I know of nothing against the
>> Webtatic repo's PHP packages.  It seems like a competent external repo
>> for CentOS/RHEL, though I have no relevant experience.  Hope that helps!
>>
>>
>> ___
>> vox-tech mailing list
>> [email protected]
>> http://lists.lugod.org/mailman/listinfo/vox-tech
> 
> ___
> vox-tech mailing list
> [email protected]
> http://lists.lugod.org/mailman/listinfo/vox-tech

___
vox-tech mailing list
[email protected]
http://lists.lugod.org/mailman/listinfo/vox-tech


Re: [vox-tech] Risks of upgrading past CentOS 6 supported PHP 5.4?

2016-06-02 Thread Rick Moen
Quoting Dr. Larry Ozeran ([email protected]):

> Since we are serving data that can change every few minutes, we
> can't move to static pages. Since we are providing that data to
> users from multiple originating sources, we pretty much have to be
> internet-facing. We have put security procedures in place, but I
> know that security is more an ongoing process than an endpoint and
> there is always more that will need to be done. If there is a better
> way to meet the needs of users other than MySQL+PHP, I am always
> open to new ideas.

Meaning no criticism, I notice in looking upthread
(http://lists.lugod.org/pipermail/vox-tech/2016-May/017013.html) that 
you mention only that your use-case involves PHP-served pages, but not
what drives that particular choice of software.

Sometimes, a local site uses PHP because it runs developed software
resting on the PHP interpreter, e.g. Wordpress, MediaWiki, etc.  
Other times, that choice resulted from 'Data for each page must be
pulled on a per-visit basis from MySQL, therefore some HTTP-invoked
process must do a SQL query and assemble page contents and we happened
to use PHP to do that because our Web guy knew how to do that.'  And
I'm sure there are other scenarios -- but dynamic is not synomyous with
PHP in any event.

Irrespective of how you arrived at that choice, obviously you would not
lightly decide to rearchitect.

A number of guides to tigthening PHP security to reduce risk exist and
may be useful.  My own modest effort, last updated when PHP5 was new, is
here:  'PHP Security' on http://linuxmafia.com/kb/Web/ .

___
vox-tech mailing list
[email protected]
http://lists.lugod.org/mailman/listinfo/vox-tech


Re: [vox-tech] Risks of upgrading past CentOS 6 supported PHP 5.4?

2016-06-02 Thread Dr. Larry Ozeran

Thanks Rick. Good information is always appreciated.

Since we are serving data that can change every few minutes, we can't 
move to static pages. Since we are providing that data to users from 
multiple originating sources, we pretty much have to be internet-facing. 
We have put security procedures in place, but I know that security is 
more an ongoing process than an endpoint and there is always more that 
will need to be done. If there is a better way to meet the needs of 
users other than MySQL+PHP, I am always open to new ideas.


Thanks,

Dr. Larry Ozeran
President, Clinical Informatics, Inc.
(530) 671-9244

On 6/2/2016 00:41, Rick Moen wrote:

Quoting Bill Broadley ([email protected]):


Does anyone know any downsides to using the webtatic PHP packages on
CentOS 6?

I've seen many machines with ugly configurations related to cpanel,
custom php installs (sometimes more than one), and fragile very hard
to reproduce apache configurations.

Although I guess I shouldn't complain, they get hacked and I get consulting.

(Sadly, this won't answer Dr. Ozeran's question, either:)

I've lately come to the conclusion, from many years as a Linux server
admin, that PHP is tolerable on a Unix machine _provided_ it isn't ever
exposed to public networks, because the ongoing security nightmare is
otherwise not justifiable.  I mean, yes if management is paying you to
do it and the money's good, but if you're the boss, say 'Hell no.'

So, e.g., every Web page on my linuxmafia.com server that used to be
dymanically assembled by the PHP interpreter at Apache page-load time
are (more recently) instead built on-disk in advance using automake or a
cron script.  Fortunately, none of those pages needed to _actually_ be
dynamic; it was just coder laziness that chose that implementation.

For example, the coder who helped me convert BALE
(http://linuxmafia.com/bale/) from its original mid-1990s static HTML
incarnation to PHP + MySQL set it up so every page load assembles the
page anew, from several PHP fragments plus the results of a MySQL query
(furnishing the events rows).  When I realised the underlying reality of
this being static data changing only once on the 1st of each month, I
converted it into a static HTML page generated by a cron job in
/etc/cron.monthly/ , and then Apache serves up just that static file.

Whole huge categories of security threat have completely away for good,
when I ditched runtime PHP.

If I had any Web applications that actually relied on the PHP
interpreter at load time, I'd try really hard to ditch them.  It really
is IMO that bad.

And I say that because, so to speak, Ranum is my guru:
http://www.ranum.com/security/computer_security/editorials/master-tzu/

That having been said:  Dr. Ozeran, I know of nothing against the
Webtatic repo's PHP packages.  It seems like a competent external repo
for CentOS/RHEL, though I have no relevant experience.  Hope that helps!


___
vox-tech mailing list
[email protected]
http://lists.lugod.org/mailman/listinfo/vox-tech


___
vox-tech mailing list
[email protected]
http://lists.lugod.org/mailman/listinfo/vox-tech


Re: [vox-tech] Risks of upgrading past CentOS 6 supported PHP 5.4?

2016-06-02 Thread Rick Moen
Quoting Bill Broadley ([email protected]):

> >Does anyone know any downsides to using the webtatic PHP packages on
> >CentOS 6?
> 
> I've seen many machines with ugly configurations related to cpanel,
> custom php installs (sometimes more than one), and fragile very hard
> to reproduce apache configurations.
> 
> Although I guess I shouldn't complain, they get hacked and I get consulting.

(Sadly, this won't answer Dr. Ozeran's question, either:)

I've lately come to the conclusion, from many years as a Linux server
admin, that PHP is tolerable on a Unix machine _provided_ it isn't ever
exposed to public networks, because the ongoing security nightmare is
otherwise not justifiable.  I mean, yes if management is paying you to
do it and the money's good, but if you're the boss, say 'Hell no.'

So, e.g., every Web page on my linuxmafia.com server that used to be
dymanically assembled by the PHP interpreter at Apache page-load time
are (more recently) instead built on-disk in advance using automake or a
cron script.  Fortunately, none of those pages needed to _actually_ be
dynamic; it was just coder laziness that chose that implementation. 

For example, the coder who helped me convert BALE
(http://linuxmafia.com/bale/) from its original mid-1990s static HTML
incarnation to PHP + MySQL set it up so every page load assembles the
page anew, from several PHP fragments plus the results of a MySQL query
(furnishing the events rows).  When I realised the underlying reality of
this being static data changing only once on the 1st of each month, I
converted it into a static HTML page generated by a cron job in
/etc/cron.monthly/ , and then Apache serves up just that static file.

Whole huge categories of security threat have completely away for good,
when I ditched runtime PHP.

If I had any Web applications that actually relied on the PHP
interpreter at load time, I'd try really hard to ditch them.  It really
is IMO that bad.

And I say that because, so to speak, Ranum is my guru:
http://www.ranum.com/security/computer_security/editorials/master-tzu/

That having been said:  Dr. Ozeran, I know of nothing against the
Webtatic repo's PHP packages.  It seems like a competent external repo
for CentOS/RHEL, though I have no relevant experience.  Hope that helps!


___
vox-tech mailing list
[email protected]
http://lists.lugod.org/mailman/listinfo/vox-tech


Re: [vox-tech] Risks of upgrading past CentOS 6 supported PHP 5.4?

2016-06-01 Thread Dr. Larry Ozeran
Thanks Bill. That's helpful. It is a physical server. You have given me 
several things to consider.


Dr. Larry Ozeran
President, Clinical Informatics, Inc.
(530) 671-9244

On 6/1/2016 21:51, Bill Broadley wrote:

On 05/18/2016 08:58 PM, Dr. Larry Ozeran wrote:

I am Googled out (repeatedly finding the same sites) and the impressive
expertise on this list gives me high hopes that someone here has an
answer or a reference.

I have found many sites talking about how to upgrade from PHP 5.4 (which
appears to be the last CentOS 6 supported version of PHP) to later
versions using webtatic. [https://webtatic.com/packages/php55/] What I
have been unable to learn are the risks of doing so.

Does anyone know any downsides to using the webtatic PHP packages on
CentOS 6?


I've seen many machines with ugly configurations related to cpanel, 
custom php installs (sometimes more than one), and fragile very hard 
to reproduce apache configurations.


Although I guess I shouldn't complain, they get hacked and I get 
consulting.


The main problem being that OS packages are upgraded regularly, if you 
build it yourself you really should be tracking the patches/upgrades 
for anything you compile your self daily or at least weekly.  In most 
environment that just isn't practical.


Normally I'd just recommend upgrading the OS when a major package is 
too old.  Sadly upgrading to CentOS 7 doesn't get you newer than PHP 
5.4.16.  Pretty sad that CentOS 7 has a PHP from 2012. Ubuntu 14.04 
LTS includes 5.5.


Ubuntu 16.04 LTS include PHP 7.0.4.

So supporting a custom PHP or doing a major OS switch to a Debian (or 
derivative) could be a difficult decision.


I do suggest a full backup before you do anything, it could be quite 
painful if things go awry.  You didn't mention if it was a virtual 
server or physical.  If virtual it could be a fairly quick experiment 
to try any of your options without breaking the original.


Sorry, no easy answer from me.

___
vox-tech mailing list
[email protected]
http://lists.lugod.org/mailman/listinfo/vox-tech


___
vox-tech mailing list
[email protected]
http://lists.lugod.org/mailman/listinfo/vox-tech


Re: [vox-tech] Risks of upgrading past CentOS 6 supported PHP 5.4?

2016-06-01 Thread Bill Broadley

On 05/18/2016 08:58 PM, Dr. Larry Ozeran wrote:

I am Googled out (repeatedly finding the same sites) and the impressive
expertise on this list gives me high hopes that someone here has an
answer or a reference.

I have found many sites talking about how to upgrade from PHP 5.4 (which
appears to be the last CentOS 6 supported version of PHP) to later
versions using webtatic. [https://webtatic.com/packages/php55/] What I
have been unable to learn are the risks of doing so.

Does anyone know any downsides to using the webtatic PHP packages on
CentOS 6?


I've seen many machines with ugly configurations related to cpanel, 
custom php installs (sometimes more than one), and fragile very hard to 
reproduce apache configurations.


Although I guess I shouldn't complain, they get hacked and I get consulting.

The main problem being that OS packages are upgraded regularly, if you 
build it yourself you really should be tracking the patches/upgrades for 
anything you compile your self daily or at least weekly.  In most 
environment that just isn't practical.


Normally I'd just recommend upgrading the OS when a major package is too 
old.  Sadly upgrading to CentOS 7 doesn't get you newer than PHP 5.4.16. 
 Pretty sad that CentOS 7 has a PHP from 2012.  Ubuntu 14.04 LTS 
includes 5.5.


Ubuntu 16.04 LTS include PHP 7.0.4.

So supporting a custom PHP or doing a major OS switch to a Debian (or 
derivative) could be a difficult decision.


I do suggest a full backup before you do anything, it could be quite 
painful if things go awry.  You didn't mention if it was a virtual 
server or physical.  If virtual it could be a fairly quick experiment to 
try any of your options without breaking the original.


Sorry, no easy answer from me.

___
vox-tech mailing list
[email protected]
http://lists.lugod.org/mailman/listinfo/vox-tech