Memory on CSS 11000

I have three different CSS 11050 switches, all with 128MB ram. all three report right around 20MB free memory, yet one on the 3 is doing much less than the other two. Is this too low of memory to be playing with or is it normal for the CSS to report this usage? Can one even still get memory upgrades for the CSS switches?

the CSS takes most of the memory to create Flow Control Block [FCB] which are then allocated to create flow connections.
From llama mode you can do a 'flow stat' to see how many free FCB you have left.
This is more important than the memory.
A device with no config and not attached to any interface would still consume most of its memory to create the FCB.
Gilles.

Similar Messages

CSS 11000 series with load balancing for high availablity servers

Hi,
We have arrived at a designed,which has 2 PIX firewall(525)in a fault tolerant mode.Two interfaces from each of the PIX is connected directly to two css1000 switch which works in a fault tolerant mode.
other Two interface from each of the two css11000 switch is connected directly to two catalys 4003 switches to which we have Servers attached.
Pls let me know whether this design will work or If you could suggest some other design for high availablity servers which are kept at the DMZ.
Pls note in our design the internet user first hit the frewall and then the CSS11000 and then catalys 4003 and finally the Servers.
If anyone can help me out,pls send me a mail at [email protected]

Are the PIXes to be setup in a failover bundle?
If you use two interfaces in each PIX to connect
to the two CSS systems, are you not complicating the
setup? Which of the two should be called the inside?
I cannot recall whether you can have two inside legs at the current PIX level.

CSS 11000 Clients' Src IP replaced with VIP's IP

Dear all,
Topology:
10.1.101.X/24 ---CSS11000--- 10.1.103.X/24(Server Farm)
Configuration:
########### Service ###########
service LBS_AP-11:11111
ip address 10.1.103.11
keepalive type tcp
keepalive port 11111
port 11111
active
service LBS_AP-12:11111
ip address 10.1.103.12
keepalive type tcp
keepalive port 11111
port 11111
active
########### Owner ###########
owner Customer-6
content LBS_AP-10:11111
vip address 10.1.103.10
add service LBS_AP-11:11111
add service LBS_AP-12:11111
port 11111
protocol tcp
active
Problem:
We notice that the traffic's Src IP address from the Client side(10.1.101.X)has been replaced by the CSS's VIP address = 10.1.103.10 (through Sniffer)
Is that correct or reasonable ?
The configuration was implemented by another Cisco Partner.
Our customer wants me to find out why they cannot see the Client's IP address as the Src IP address when the traffic is from 101.X to 103.X
Is there still anything that I have to provide to you experts, plz let me know this.
Appreciate ^__^
CSS# show version
Version: ap0610107a (6.10 Build 107)
Flash (Locked): 6.10 Build 107
Flash (Operational): 6.10 Build 107
Type: PRIMARY
Licensed Cmd Set(s): Standard Feature Set

Dear Gilles,
Yes, we do have a group statement as follow:
########## Group ##########
group LBS
vip address 10.1.103.10
add destination service LBS_AP-11:11111
add destination service LBS_AP-12:11111
active
The question is . . . when we remove the group statement as mentioned above, the patron can finally see the traffic definitely from the client side(101.X/24) and destined to the VIP(10.1.103.10).
But, unfortunately, there are some DB servers on the same side of the Server Farm(103.X) having no access to the VIP(10.1.103.10).
(They could & did access to the VIP when the group statment was configured.)
What can I do now for this situation?
I am totally confused with "sourcegroup", "one-armed scenario",& "ACL".
The patron said they want to see the true Src IP addresses from 101.X and still have the DBs on the 103.X network access to the VIP 10.1.103.10.
DBs IP Addresses range from 10.1.103.15 to 10.1.103.20, etc.
If there is still anything that I should provide to you experts, plz let me know that.
Appreciate ^__^
########## Relevant Configuration ##########
circuit VLAN101
redundancy
description "SC-SAA-3548 VLAN101"
ip address 10.1.101.30 255.255.255.0
circuit VLAN103
redundancy
description "SC-SAA-3548 VLAN103"
ip address 10.1.103.254 255.255.255.0
acl 20
clause 10 permit tcp any destination any
clause 20 permit any any destination any
apply circuit-(VLAN101)
apply circuit-(VLAN103)
Best Regards
Sincerely,

What is the appropriate product name for CSS 11500 on Bug Toolkit

Today I tried to search DDTs of CSS 11500 on Bug Toolkit (http://www.cisco.com/cgi-bin/Support/Bugtool/home.pl), however I can not find out appropriate product name corresponding to CSS 11500.
Before I had searched DDTs of CSS 11500 on Bug Toolkit many times, at that time, if my memory correct..
I selected "Cisco CSS 11500 Series Content Services Switches" in the list of "Search for bugs in other Cisco software and hardware products" on Bug Toolkit.
But I can not find this product name today.
Do you know what product name appropriate for CSS 11500 on Bug Toolkit ?
Your information would be appreciated.
Best regards,

Hi Gilles,
Thank you for your cooperation.
Today, I can find the CSS at "new Bug Toolkit".
http://tools.cisco.com/Support/BugToolKit/action.do?hdnAction=searchBugs
Select Product Category: Application Networking Services
Select Product: Cisco CSS 11000 Series Content Services
So I understand I should go "new Bug Toolkit" instead of old "Bug Toolkit" to
search any DDTs for CSS 11500.
Many thanks.
Best regards,

CSS cookieurl: case sensitive?

Hello,
As far as I understand, the HTTP header cookie field, and the embedded cookie (in a URL) are case-sensitive. Is this correct?
The settings look like this:
content stickyCookie
advanced-balance cookieurl
string prefix "SESSION_ID="
On the HTTP header the cookie field looks like this:
Set Cookie: SESSION_ID=ABCDEF
But if instead of this, the request comes from a client with the cookie embedded in the url, it looks like this:
http://localhost/Info/index;session_id=ABCDEF?page=home.
Is this field case-sensitive for the CSS? I know it is for the ACE, as I have done this and there is the "cookie secondary" command where you can specify how the url-embedded cookie looks like.
So I am afraid there is no way of making this work on a CSS?
Any help will be much appreciated!

Hi Javier,
The Cisco CSS is case-sensitive when searching for this string.
When configuring the CSS 11000 and 11500, determine first if you need a server cookie string. If the string operation under the content rule is match-service-cookie, which is the default setting, this parameter must be configured. In this case the service cookie string is matched against the cookie contained in the HTTP header for load balancing decision.
If the string operation under the content rule is set to the hash method, this parameter is not needed. In order to select one of the available servers, the cookie contained in the HTTP header is mathematically processed using a hashing algorithm. If available, the hash algorithm assigna the connection carrying a certain cookie to the same server. This example focuses on the match-service-cookie case.
Note: Cookies are case sensitive.
Configuring the Service
training4(config># service server_g
training4(config-service[server_g])# string LV2KJK (the server cookie text)
training4(config># service server_h
training4(config-service[server_h])# string AARIKA
Configuring the Content Rule
Note: Cookies require a Layer 5 rule.
You can create a Layer 5 rule by adding a URL.
For example, .
A Layer 4 rule can be promoted to a Layer 5 rule by issuing the advanced-balance cookies command.
Choose the advanced-balance method.
training4(config-owner-content[cookie-layer5])# advanced-balance cookies
Configure the string operation.
training4(config-owner-content[cookie-layer5])# string operation ?
match-service-cookie (DEFAULT)
hash-crc32
hash-xor
hash-a
Define the starting/ending bytes.
training4(config-owner-content[cookie-layer5])# string range 1 to 200
Start byte position of cookie/url after header (Range:1-600)
Specify the prefix located in the string range.
training4(config-owner-content[cookie-layer5])# string prefix "ASPSESSION"
"Quoted textual information"(Len: 0-32)
Indicate how many bytes to skip after the starting prefix.
training4(config-owner-content[cookie-layer5])# string skip-length 9
"Quoted textual information"(Len: 0-32)
Indicate how many bytes after the prefix/skip-length make-up the string.
training4(config-owner-content[cookie-layer5])# string process-length 6
Integer value(Range: 0-64)
If no string process length is configured, search after end of string character.
training4(config-owner-content[cookie-layer5])# string eos-char "&"
"Quoted textual information"(Len: 0-5)
Specify the failover in the event that a server goes down or is suspended.
training4(config-owner-content[cookie-layer5])# sticky-serverdown-failoversticky-srcip
sticky-srcip-dstport
sticky-srcip
balance (Default)
redirect
reject
Below is an example of a cookie string, and how some of the parameters work.
ASPSESSIONJJKKJJKK=LV2KJK44444444
!--- The string prefix = ASPSESSION.
!--- The string skip-length = 9. Skip
!--- nine characters after the prefix.
!--- The string process-length = 6, which would make
!--- the string LV2KJK matching service server_g.
ASPSESSIONSQPMMJHK=AARIKAMDESLD
!--- Matches service server_h.
Below is a sample configuration for the cookie strings.
!************************** SERVICE **************************
service server_g
ip address 172.17.63.240
string LV2KJK
active
service server_h
ip address 172.17.63.241
string AARIKA
active
!*************************** OWNER ***************************
owner braden
content server-cookie
protocol tcp
vip address 172.17.63.199
port 80
advanced-balance cookies
string range 1 to 200
string prefix "ASPSESSION"
string skip-length 9
string process-length 6
add service server_g
add service server_h
active
Kindly tell if this information is of any use to you.
Sachin garg

Servers behind CSS Browse the Internet

Can anyone show me how to configure on CSS for servers behind CSS 11000 browse the Internet ? I am using NAT.
thanks in advance
Andy

if you need the CSS to nat the servers ip addresses when they go out to the Internet, all you need is a group.
ie:
service server1
ip 10.x.x.x
active
servoce server2
ip 10.x.x.x
active
group natservers
vip address x.x.x.x
add service server1
add service server2
active
If you already had a group using these services, you will need something a little more tricky.
Just start with the above and let us know if it works.
Gilles.

CSS 11051: Sorry Server receives request although the normal server is up

Hello,
my customer has configured a sorry for his server. If the normal server is down the Sorry Server receives the requests. That works fine. But if the normal server comes back the Sorry Server still receives some requests( 2 hours and more). Has anybody an idea what might be the reason for that ?
regards
Dietrich Schleyer
content webserver
add service server12
vip address 10.40.52.20
primarySorryServer server13
protocol tcp
port 80
url "/*"
no persistent
active
service server12
ip address 10.40.52.12
port 80
protocol tcp
keepalive type named applicationwww01
active
service server13
ip address 10.40.52.13
protocol tcp
port 80
keepalive type named applicationwww02
active
keepalive applicationwww01
ip address 10.40.52.12
port 80
type http non-persistent
uri "/test.html"
frequency 10
method get
active
keepalive applicationwww02
ip address 10.40.52.13
port 80
uri "/test.html"
frequency 10
method get
type http non-persistent
active

According to: http://www.cisco.com/warp/public/117/css_sorry_server.html After the CSS 11000 directs requests to a primary sorry server, the switch will continue to use the primary sorry server even when the original server becomes functional. To force the connection back to the original server, you must suspend the primary sorry server or wait until the connection is dropped or times out. When a new session is initiated by the CSS 11000, the connection should go back to the original server.

CSS polling and response from servers

Hello,
Is it possible to poll webservers from CSS for a specific content on the webpage.
At present we poll for a URL. what we want to ensure if the web page is modified then CSS should take server out of farm.
Also all these web pages have drop down box and connects to a DB server when a select is made. whats the best practice to implement if these drop down doesnt work , where the web server lost the connection DB server the CSS no longer sends the request to that server
Thanks

Hi,
The CSS 11000 HTTP keepalive uses one of two available methods:
HEAD
GET
They are specified by using the method command while in the keepalive configuration mode (config-keepalive).
When an HTTP HEAD keepalive is issued in the CSS 11000, a 200 Status OK is expected in the response packet from the server it is querying. If the 200 OK status is not returned, the CSS 11000 considers the service down. If the CSS 11000 receives any other status code such as a 404 Object Not Found or a 302 Object Moved, the CSS 11000 will view this as an incorrect response and mark this service as down. The head method is the default. The CSS 11000 does not compute a reference hash value for this type of keepalive.
When an HTTP GET keepalive is used, the CSS 11000 expects not only to see the 200 Status OK in the response packet, but it will also run a checksum on the entity body. The first time the CSS 11000 goes out to query, the server will run a checksum on the entity body and store that hash value for future queries. If future responses pass back a 200 Status OK and the hash value that is calculated is different from the stored value, then the service is considered down. If the 200 Staus OK is not returned, or if the 200 OK status is returned but the hash value is different from the reference hash value, the CSS 11000 considers the service down.
When you specify an URL for an HTTP keepalive, the CSS 11000 calculates a hash value for the Web page specified in the URL. If the Web page changes, the hash value no longer matches the original hash value, and the CSS 11000 assumes that the service is down. To prevent the CSS 11000 from assuming that a service is down because of a hash value mismatch, specify the keepalive method as the head. Because of the nature of HTTP GET keepalives and the way they are calculated, do not use them on dynamically changing pages because their checksums change constantly.
With CSS if you use GET method in a keepalive, It hashes the page and if the subsequent GET requests gives a different hash the service is marked down.
Here is an example:
service WEB01
ip address 192.168.11.115
keepalive method get
keepalive type http
keepalive uri "/General/GenAppAlive.asp?ControlAction=InitialLoad"
active
Sachin Garg

CSS URL vs. F5 Big-IP

Can anyone tell me the difference in performance between CSS URL switching and F5 Big-IP?
TIA

Murabi,
You will what to look at http 1.1 performance numbers not just the number of connections you can handle. See the follow url for the CSS 11000 independant performance testing http://www.zdnet.com/etestinglabs/stories/main/0,8829,2462091,00.html

How to setup CSS as reverse proxy server without cache server

Hi, question regarding CSS capabilities.
Is there a way to setup CSS (11000 series or 11500) to act as a reverse proxy server without caching server in place? Or is it possible for CSS to cache contents on its own?
And what is disk space on CSS used for?
Thank you.

Thank you for clarification on disk usage on CSS. I have gone through sample config for reverse proxy caching
http://www.cisco.com/warp/customer/117/CSS_CEreverseproxy.html
Does this mean only way to have CSS to do reverse proxying is to have cache engine in conjunction with CSS?
Thank you again in advance.

IMac 2010 cannot afford to play CSS properly.

I recently had issues with my iMac 27" 2010 model and got the logic board and the hard drives replaced.
My iMac arrived last week. Since then (and even before), I couldn't play CSS properly; let alone Portal 2 and other stuff.
When I pickup a sniper, snipe once and move, it lags: a very unpleasant experience. When I snipe twice and am just still, it feels ridiculous.
I ran tech tool and everything seems to be fine. The GPU uses half of the memory in CSS.
I don't play much games. CSS is the only game I play as I use this machine for development reasons.
Any one with an opinion on this?

Thanks for the reply.
I did mention CSS - Counter Strike Source which is a very old game.
I understand the situation with less RAM, but I used to play this game back in the day on probably 2gigs of RAM? But that was on Windows though.
I don't think CSS is a powerpc game. It now is sold through Steam[client].
So I did buy 8gigs of RAM last week which I put in this iMac in the evening. The overall system seems to be a bit fast. But really, the game is still the same.
I don't know what's gone wrong.

CSS 11051 Stateful Failover

We have received a note concerning stateful failover of the CSS series of products, where the CSS 110XX series doesn't support stateful failover, however the CSS 115XXX will. Here is the digest of the message;
On 3/6, Hosting Engineering and Operations issued an alert regarding the
CSS 11000 load balancer. This is an update to that alert.
Since that time, we have experienced another hardware failure of this
model device.
In response to this situation, the following has occured:
* Platform Engineeringis in the process of removing the CSS 11000
from the SOE. Itis on target to be removed in April.
* Operations hasre-inforced our escalation procedures with CISCO.
Qwest is to beissued a RMA immediately for this model.
* For newconfigurations including a CSS 11000, CCAR will require
an Individual CaseBasis (ICB) review and approval.
* For existing premiumand above customers whoes configurations
include a CSS 11000, HostingOperations is planning to replace them with
compatible device. These changes have been pre-approved by CCAR as long
as:
* the networktopology remains the same
* redundancy ispreserved
* CCAR gets notifiedof the replacement model so we can update our
records
* For existing basicand enhanced customers, we are drafting a
communique that alerts them tothe performance issues exerienced by Qwest
and providing suggestedalternative solutions.
In response to recent questions from the field.....
Stateful failover with redundant CSS 11000 Series Load Balancers:
The Bottom Line: Cisco CSS 11000 Series Load Balancers do not support
stateful
failover.
Will Cisco ever support this?: Yes, this is supported in the CSS 11500
Series,
known as Adaptive Session Redundancy (ASR)
I need this today, what can I do?: Choose an alternative product. The
F5 BIG-IP
load balancers support this functionality.
What is stateful failover anyhow?
Stateful failover is a technology that can maintain state information
between
the active load-balancer and the standby load-balancer. This state
information
can include: persistence mapping, telnet sessions, ftp sessions, tcp
session
state, etc...
Why should I be concerned?
Without state synchronization applications can break if there is a
failover from
the active to standby unit. FTP Sessions will be broken, Telnet
sessions will
be broken, and most importantly persistence state mapping will be lost.
What do I need to listen for to determine if stateful failover is
important?
1. E-commerce applications that require persistence mapping.
Persistence
mapping will keep a client session mapped to the same server for a
specified
amount of time. This is often important with shopping cart and other
e-commerce
applications.
2. Long-lived sessions. Whether they are planning to transfer large
files via
FTP or long-lived telnet sessions. Anytime a connection will be
required for a
long time and starting over is not an acceptable condition, then
stateful
failover is important.
Does this sound correct or is this a bunch of hot air?

Yes. Stateful failover, or ASR as it is sometimes called, is available on the CSS 11500 and Catalyst 6500 Content Switch Modules (CSM) load balancing platforms. It is not supported on the CSS 11000 due to architectural limitations of that platform.
Stateful failover is available on these Cisco platforms today.
mikep

Can anyone tell me the difference in performance between CSS URL switching

first
Can anyone tell me the difference in performance between CSS URL switching and F5 Big-IP?
second
Can anyone tell me the difference in performance between CSS URL switching and alteon ?
third what is best overally?
i think alteon is best
is that right?
best regard

It looks like the primary question here is performance, in which case:Performance is not an issue, so long is it is sufficient. In the case of CSS 11000 and Alteon, performance fall within the same order of magnitude (supporting web sites with several billion hits per day.) F5 does not have sufficient performance, due to platform and OS limitations (Ive heard as low as 50 connections per second in complex configurations). The Cisco CSM posts up over 10x the performance at 200k flows/second.
Typically the primary concern is features, then CSS 11000 switches lead with a wide and flexible array of features that are not only helpful to network and web administrators, but well integrated too. CSS 11000 switches offer configuration through CLI, Web GUI, and XML. CSS 11000 collects statistics that can be exported to non-Cisco applications for billing and management.
The CSS11000 also supports URL load balancing and HTTP header balancing within one content rule, with complex matching. Further, it supports user agent, pragma/no-cache, host field, cookie field, language field, accept, accept charset, accept-encoding, and Connection within the http header field.
In addition, the CSS11000 matches up to 128 bytes with support of wildcards anywhere within the string.On the other hand, Nortel (Alteon) - HTTP header load balancing is not supported on the same VIP as URL load balancing. Eg. This means that a simulated WAP user cannot be directed to a server while load balancing "normal" browsers to their servers based on URL without using two separate VIPs. Only ONE http header load balance is supported on the entire switch. This limits you to either User Agent (WAP, Netscape, IE, Palm, etc), OR pragma/no cache (do not send the user to cache, allow (the user) to go to the origin), or Host field (allowing you to direct on domain name), or Cookie. Also, Nortel does not support the language field.
In regards to F5, many of the performance claims are based on HTTP 1.0 requests (most web sites today are not using HTTP 1.0), Many emerging applications rely on HTTP 1.1 rather than HTTP 1.0. Also, the BIG-IP cannot spoof the connection to detect the URL, cannot do NAT on the flow, and cannot maintain states for persistent connections.
Overall, I think CSS switches are the lowest cost to own, and most effective of all the load balancing platforms on the market.

CSS redundant interface and DNS server

We're attempting to implement a pair of CSS's using redundant ASR and GSLB where the CSS's act as DNS servers.
But I'm not sure if the 2 features are compatible. The CSS's answer DNS queries to their direct interface but not the redundant interface.
Does anyone have any suggestions or work-arounds? We're running version 8.20.
TIA,
Dan

Dan doing some research I can see that the option to configure redundant-interface to resolve dns queries is not included on CSS 11500 series, this from the documentation.
On the document for CSS 11000 series that I provided before shows:
Configuration Requirements and Restrictions
The following requirements and restrictions apply to the configuration of this feature.
â¢You can configure this feature only on Cisco 11000 series CSSs (not 11500)
If I look at the redundant-interface configuration on old CSS 11000 series I see the option for dns:
http://www.cisco.com/en/US/docs/app_ntwk_services/data_center_app_services/css11000series/v6.10/configuration/advanced/guide/VIPRedun.html#wp1067528
Look at this line:
dns-server - Keyword that enables the CSS to respond to DNS queries destined for the redundant interface IP address. For more information, see the "Configuring a Redundant Virtual Interface to Respond to DNS Requests" section.
On new CSS 11500 series this option is not available:
http://www.cisco.com/en/US/docs/app_ntwk_services/data_center_app_services/css11500series/v8.20_v8.10/configuration/redundancy/guide/VIPRedun.html#wp1067528
I am trying to find if there is any workaround but so far semms that is expected to miss this feature on CSS11500.

CSS Console Cable

I have a Cisco CSS 11001 that's been sitting around for a while and I'm trying to console into it. I don't have the original cable that came with it and using a standard molded Cisco console cable isn't working. I've checked the manual online and the terminal settings are correct, so I can only assume it's the cable. Any ideas?

Hi,
The CSS 11000 series need a different cable, this link explains how to build it:
http://www.cisco.com/en/US/products/hw/contnetw/ps789/products_tech_note09186a00800a3f28.shtml
Also make sure to set Flow Control to none on the port settings.
Hope it helps!!

Memory on CSS 11000

Similar Messages

Maybe you are looking for