Re: [Zenloadbalancer-support] http farm - downlading files 2GB

2015-03-02 Thread Mathieu Chateau
found this on pound. 2GB limit is fixed in pound 2.6e apparently:

*ANNOUNCE: Pound - reverse proxy and load balancer - v2.6e*
Robert Segall roseg(at)apsis.ch*2011-04-25 11:16:44*[ FULL
http://www.apsis.ch/pound/pound_list/archive/2011/2011-04/1302530799000/index_html?fullMode=1#1303723004000
 ]

This is to announce the release of Pound v2.6e. This is an experimental
version - the fifth (and hopefully the last prior to the stable
release) in the 2.6 series. Changes since version 2.6d:

Bug fixes:
- fixed problem in SNI certificate storage
- changed long to long long for support of requests larger than 2GB

The software is at version 2.6e (beta quality). Further testing
(especially under heavy loads), improvements and suggestions are
welcome.[...]


Cordialement,
Mathieu CHATEAU
http://www.lotp.fr

2015-03-03 8:15 GMT+01:00 Mathieu Chateau mathieu.chat...@lotp.fr:

 Hello,

 On one of my website, i have a 18GB file. When downloading it, it always
 stop at 2GB.
 I use the last version of chrome, and for sure it can download files
 bigger than 2GB, which was a old limit

 If I do it directly on the web server (skipping zen) it works beyond 2GB

 In Zen load balancer, i guess it's the pound process which has the issue
 as it's a 32 bit binary?

 /usr/local/zenloadbalancer/app/pound/sbin# file pound
 pound: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV),
 dynamically linked (uses shared libs), for GNU/Linux 2.6.18, not stripped


 regards,
 Mathieu CHATEAU
 http://www.lotp.fr

--
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the 
conversation now. http://goparallel.sourceforge.net/___
Zenloadbalancer-support mailing list
Zenloadbalancer-support@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/zenloadbalancer-support


[Zenloadbalancer-support] http farm - downlading files 2GB

2015-03-02 Thread Mathieu Chateau
Hello,

On one of my website, i have a 18GB file. When downloading it, it always
stop at 2GB.
I use the last version of chrome, and for sure it can download files bigger
than 2GB, which was a old limit

If I do it directly on the web server (skipping zen) it works beyond 2GB

In Zen load balancer, i guess it's the pound process which has the issue as
it's a 32 bit binary?

/usr/local/zenloadbalancer/app/pound/sbin# file pound
pound: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV),
dynamically linked (uses shared libs), for GNU/Linux 2.6.18, not stripped


regards,
Mathieu CHATEAU
http://www.lotp.fr
--
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the 
conversation now. http://goparallel.sourceforge.net/___
Zenloadbalancer-support mailing list
Zenloadbalancer-support@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/zenloadbalancer-support


Re: [Zenloadbalancer-support] Ghost

2015-03-02 Thread Genoint G
I sent the following message on 2/17 with no response.  What are the 
remediation/patching steps to fix Ghost with Zen LB 3.03?  regards, John
 

 On Tuesday, February 17, 2015 1:47 PM, Genoint G geno...@yahoo.com wrote:
   

  What are the remediation steps for Ghost on Zen Load Balancer version 3.03 
and 2.0?  I see that there is a patch released?  I know 2.0 is quite old, if we 
need to we can upgrade.
Regards, John


   --
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the 
conversation now. http://goparallel.sourceforge.net/___
Zenloadbalancer-support mailing list
Zenloadbalancer-support@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/zenloadbalancer-support


Re: [Zenloadbalancer-support] Load Balancer on other WAN side not accessible

2015-03-02 Thread Cody Rose
Juniper ScreenOS and JunOS both have commands to allow a gratuitous arp
to update the local arp table on the device. By default this is disabled
for security and left up to the admin to enable if needed.

http://kb.juniper.net/InfoCenter/index?page=contentid=KB23573

http://kb.juniper.net/InfoCenter/index?page=contentid=KB16215

http://www.juniper.net/documentation/en_US/junos14.2/topics/usage-guidelines/interfaces-configuring-gratuitous-arp.html

I've needed these commands for multiple HA environments where ScreenOS
or JunOS is in use.

Thanks,

Cody


On Mon, 2 Mar 2015 09:58:20 +0100, Emilio Campos
emilio.campos.mar...@gmail.com wrote:
 Mathew /etc/rc.local is executed only when the server start not when
 the cluster switches.
 
 Please Roger try to use /usr/local/zenloadbalancer/config/zlb-start
 adding a script that run 4 icmp ping packet to your gw. 
 
 It is added to the Enterprise Edition at the moment, it will be added
 in the next zen community edition. 
 
 Regards
 
 2015-03-02 8:57 GMT+01:00 Mathieu Chateau :
 
 Hello,
 
 I am new on zen, but maybe a continous ping can be added in
 /etc/rc.local or the start script in zen folder ? This may be a
 workaround waiting for the new feature
 
 Regards,
 Mathieu Chateau
 
 Envoyé de mon iPad
 
 Le 2 mars 2015 à 08:44, Roger Sikorski  a écrit :
 
 Hello Emilio,
 
  
 
 thanks for you quick replay.
 
  
 
 Are there any plans when this property will be developed?
 
  
 
 Greetings
 
 Roger 
 
  
 
 VON: Emilio Campos [mailto:emilio.campos.mar...@gmail.com [3]] 
 GESENDET: Sonntag, 1. März 2015 22:38
 AN: zenloadbalancer-support@lists.sourceforge.net [4]
 CC: Sazzad Hussein; Jonathan Aldred
 BETREFF: Re: [Zenloadbalancer-support] Load Balancer on other WAN side
 not accessible
 
  
 
 Hi, it is a know issue with some firewalls. Zen sends gratuitous arp
 packages to the broadcast when cluster switches to force the ARP table
 to be updated but some firewalls require a direct ICMP packet to
 update the ARP table, as you reported it is your case. 
 
 We are going to develope this property 'send ping packets to the gw if
 the zen switches' for future releases.
 
 Thanks for your report.
 
 El 01/03/2015 22:11, Roger Sikorski  escribió:
 
 Hello Emilio,
 
  
 
 yes Juniper is the gateway. I also forced the first 
 
  
 
 After the connection to the load balancer is not more reachable from
 the other location, the LB2 switched automatically to LB1. 
 
  
 
 I tried then a ping from both load balancer to the Gateway. The both
 pings was continuous possible.
 
 After the pings the we got an email from our monitoring system that
 the load balancer is from the other location again accessible. 
 
  
 
 Between the first e-mail that the load balancer wasn’t reachable and
 now reachable is (1 minute later after the ping) are 43 minutes.
 
  
 
 I didn’t forced the second load balancer to switch to the first load
 balancer.
 
 This was tested with the option force the second load balancer as a
 master.
 
  
 
 I activated now the maintenance mode and will do the tests ask you
 asked and inform you.
 
  
 
 Greetings
 
 Roger 
 
  
 
 VON: Emilio Campos [mailto:emilio.campos.mar...@gmail.com [6]] 
 GESENDET: Montag, 23. Februar 2015 14:48
 AN: zenloadbalancer-support@lists.sourceforge.net [7]
 CC: Sazzad Hussein; Jonathan Aldred
 BETREFF: Re: [Zenloadbalancer-support] Load Balancer on other WAN side
 not accessible
 
  
 
 Hi Roger, sorry for the late response, I have been checking your logs
 and I have detected your cluster service is switching, I would require
 you confirm that you problema appears when the cluster service
 switches so could you run some test forcing the master node to
 maintenance?  
 
  
 
 Once you confirm you lose the access to the service from juniper ( I
 guess juniper is your gw),  I would require you run ping from your
 active zen server (node2) to the gw (juniper), and confirm if one you
 run a continuous ping the problem is solved..
 
  
 
  
 
  
 
 2015-02-23 10:56 GMT+01:00 Roger Sikorski :
 
 Hello Emilio,
 
  
 
 did you found something out?
 
  
 
 I shut down the second load balancer (for the cluster) and we got no
 more notification from our monitoring system that the load balancer
 isn’t more reachable.
 
 Maybe that helps you.
 
  
 
 Greetings
 
 Roger Sikorski
 
  
 
 VON: Roger Sikorski [mailto:roger.sikor...@de.rr-icecream.eu [9]] 
 GESENDET: Donnerstag, 19. Februar 2015 13:55
 AN: zenloadbalancer-support@lists.sourceforge.net [10]
 CC: Sazzad Hussein; Jonathan Aldred
 
 BETREFF: Re: [Zenloadbalancer-support] Load Balancer on other WAN side
 not accessible
 
  
 
 Hello Emilio,
 
  
 
 did you found something out?
 
  
 
 Greetings
 
 Roger
 
  
 
 VON: Roger Sikorski [mailto:roger.sikor...@de.rr-icecream.eu [11]] 
 GESENDET: Montag, 16. Februar 2015 09:46
 AN: zenloadbalancer-support@lists.sourceforge.net [12]
 CC: Jonathan Aldred
 BETREFF: Re: [Zenloadbalancer-support] Load Balancer on other WAN side
 not accessible
 
  
 
 

Re: [Zenloadbalancer-support] Ghost vulnerability

2015-03-02 Thread Mike Watson
Apologies - I'd asked this question a month ago, and Colin Waring kindly
responded. I've been offline for awhile and didn't see that someone else
had asked the same question.

To summarize Colin's response, you need to patch your Debian platform to
the LTS version., which is supported through Feb 2016.

Instructions are available at

https://wiki.debian.org/LTS/Using

...but as Colin notes, you also need to install the key for the LTS
repository.

Here are the steps I used, which worked well for me, and which are
summarized from the document on debian.org.

1) Edit /etc/apt/sources.list

  Add these two lines:

   deb http://http.debian.net/debian/ squeeze-lts main contrib non-free

   deb-src http://http.debian.net/debian/ squeeze-lts main contrib non-free

The security repositories that are already there are no longer needed, as
the last point release for squeeze have been done, so remove them.

So that in total /etc/apt/sources.list includes these lines:

   deb http://http.debian.net/debian/ squeeze main contrib non-free

   deb-src http://http.debian.net/debian/ squeeze main contrib non-free

   deb http://http.debian.net/debian squeeze-lts main contrib non-free

   deb-src http://http.debian.net/debian squeeze-lts main contrib non-free

2) If /etc/apt/apt.conf contains something like this:

APT::Default-Release squeeze;

then you must either comment it out, or replace it with:

APT::Default-Release squeeze-lts;



3) Install the key for LTS:

apt-get install debian-keyring debian-archive-keyring

cd /
Run apt-get update and apt-get upgrade

-- 
Mike Watson
WMC Information Services
m mmwat...@wmcnet.orgmwat...@wyomingmedicalcenter.org
307-577-2028
--
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the 
conversation now. http://goparallel.sourceforge.net/___
Zenloadbalancer-support mailing list
Zenloadbalancer-support@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/zenloadbalancer-support


Re: [Zenloadbalancer-support] Ghost

2015-03-02 Thread Genoint
Great thanks. I will have a look. 



 On Mar 2, 2015, at 12:26 PM, Colin Waring co...@dolphinict.co.uk wrote:
 
 It had already been discussed on the list which is presumably why you didn’t 
 see a response.
  
 Just dig up the list archives and search for Ghost for emails from the 
 beginning of Feb. You basically need to get your OS onto the LTS version of 
 Debian and apt-get will do the patches for you.
  
 All the best,
 Colin Waring.
  
 From: Genoint G [mailto:geno...@yahoo.com] 
 Sent: 02 March 2015 15:52
 To: zenloadbalancer-support@lists.sourceforge.net
 Subject: Re: [Zenloadbalancer-support] Ghost
  
 I sent the following message on 2/17 with no response.  What are the 
 remediation/patching steps to fix Ghost with Zen LB 3.03?  regards, John
  
  
 
 On Tuesday, February 17, 2015 1:47 PM, Genoint G geno...@yahoo.com wrote:
  
 
  What are the remediation steps for Ghost on Zen Load Balancer version 3.03 
 and 2.0?  I see that there is a patch released?  I know 2.0 is quite old, if 
 we need to we can upgrade.
  
 Regards,
 John
  
  
--
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the 
conversation now. http://goparallel.sourceforge.net/___
Zenloadbalancer-support mailing list
Zenloadbalancer-support@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/zenloadbalancer-support


Re: [Zenloadbalancer-support] Load Balancer on other WAN side not accessible

2015-03-02 Thread Mathieu Chateau
Hello,

I am new on zen, but maybe a continous ping can be added in /etc/rc.local
or the start script in zen folder ? This may be a workaround waiting for
the new feature

Regards,
Mathieu Chateau

Envoyé de mon iPad

Le 2 mars 2015 à 08:44, Roger Sikorski roger.sikor...@de.rr-icecream.eu a
écrit :

Hello Emilio,



thanks for you quick replay.



Are there any plans when this property will be developed?



Greetings

Roger



*Von:* Emilio Campos [mailto:emilio.campos.mar...@gmail.com
emilio.campos.mar...@gmail.com]
*Gesendet:* Sonntag, 1. März 2015 22:38
*An:* zenloadbalancer-support@lists.sourceforge.net
*Cc:* Sazzad Hussein; Jonathan Aldred
*Betreff:* Re: [Zenloadbalancer-support] Load Balancer on other WAN side
not accessible



Hi, it is a know issue with some firewalls. Zen sends gratuitous arp
packages to the broadcast when cluster switches to force the ARP table to
be updated but some firewalls require a direct ICMP packet to update the
ARP table, as you reported it is your case.

We are going to develope this property 'send ping packets to the gw if the
zen switches' for future releases.

Thanks for your report.

El 01/03/2015 22:11, Roger Sikorski roger.sikor...@de.rr-icecream.eu
escribió:

Hello Emilio,



yes Juniper is the gateway. I also forced the first



After the connection to the load balancer is not more reachable from the
other location, the LB2 switched automatically to LB1.



I tried then a ping from both load balancer to the Gateway. The both pings
was continuous possible.

After the pings the we got an email from our monitoring system that the
load balancer is from the other location again accessible.



Between the first e-mail that the load balancer wasn’t reachable and now
reachable is (1 minute later after the ping) are 43 minutes.



I didn’t forced the second load balancer to switch to the first load
balancer.

This was tested with the option force the second load balancer as a master.



I activated now the maintenance mode and will do the tests ask you asked
and inform you.



Greetings

Roger



*Von:* Emilio Campos [mailto:emilio.campos.mar...@gmail.com]
*Gesendet:* Montag, 23. Februar 2015 14:48
*An:* zenloadbalancer-support@lists.sourceforge.net
*Cc:* Sazzad Hussein; Jonathan Aldred
*Betreff:* Re: [Zenloadbalancer-support] Load Balancer on other WAN side
not accessible



Hi Roger, sorry for the late response, I have been checking your logs and I
have detected your cluster service is switching, I would require you
confirm that you problema appears when the cluster service switches so
could you run some test forcing the master node to maintenance?



Once you confirm you lose the access to the service from juniper ( I guess
juniper is your gw),  I would require you run ping from your active zen
server (node2) to the gw (juniper), and confirm if one you run a continuous
ping the problem is solved..







2015-02-23 10:56 GMT+01:00 Roger Sikorski roger.sikor...@de.rr-icecream.eu
:

Hello Emilio,



did you found something out?



I shut down the second load balancer (for the cluster) and we got no more
notification from our monitoring system that the load balancer isn’t more
reachable.

Maybe that helps you.



Greetings

Roger Sikorski



*Von:* Roger Sikorski [mailto:roger.sikor...@de.rr-icecream.eu]
*Gesendet:* Donnerstag, 19. Februar 2015 13:55
*An:* zenloadbalancer-support@lists.sourceforge.net
*Cc:* Sazzad Hussein; Jonathan Aldred


*Betreff:* Re: [Zenloadbalancer-support] Load Balancer on other WAN side
not accessible



Hello Emilio,



did you found something out?



Greetings

Roger



*Von:* Roger Sikorski [mailto:roger.sikor...@de.rr-icecream.eu
roger.sikor...@de.rr-icecream.eu]
*Gesendet:* Montag, 16. Februar 2015 09:46
*An:* zenloadbalancer-support@lists.sourceforge.net
*Cc:* Jonathan Aldred
*Betreff:* Re: [Zenloadbalancer-support] Load Balancer on other WAN side
not accessible



Hello,



yes we are using a zen cluster. Cluster type is ,,LB1 or LB2” can be
masters.



Logs are directly send to you.



Greetings

Roger Sikorski



*Von:* Emilio Campos [mailto:emilio.campos.mar...@gmail.com
emilio.campos.mar...@gmail.com]
*Gesendet:* Montag, 16. Februar 2015 09:15
*An:* zenloadbalancer-support@lists.sourceforge.net
*Cc:* Jonathan Aldred
*Betreff:* Re: [Zenloadbalancer-support] Load Balancer on other WAN side
not accessible



are you working with a zen cluster? anyway could you please attach logs?



/usr/local/zenloadbalancer/logs/*

/var/log/messages

/var/log/syslog



Make a ta.gz file and feel free to send me directly.



Thanks!







2015-02-13 8:53 GMT+01:00 Roger Sikorski roger.sikor...@de.rr-icecream.eu:

Hello,



today morning we got messages from our monitoring systems on other factory
location that the ZenLoad Balancer isn’t more reachable.



In the local network the Load Balancer was accessible.



We restarted for test purposes the Juniper firewall. The access from the
other factory location was then possible.



I attached a Wireshark