Application Delivery (ADX)

Reply
Occasional Contributor
Posts: 17
Registered: ‎11-10-2010

Persistent hash problem after configuring max-conn, disable/enable

I deployed ServerIron ADX4000 as cache load balancing using TCS feature. In this environment, I have 3 cache servers (SwiftCache). ServerIron has been done some TCS feature such url hashing, selective cache content base on cache server, spoofing support, http-cache-control etc.

Everything has been alright until I found hashing problem when I disable/enable, configure max-conn or add new cache server. Hashing doesn’t work correctly like result below:

telnet@SI-ADX4000-vbp#show cache-group

############ BP 1/1 output start #############

Cache-group 1 has 3 members Admin-status = Enabled Active = 0

Hash_info: Dest_mask = 255.255.255.0 Src_mask = 0.0.0.0

Cache Server Name                Admin-status Hash-distribution  L7-Hash-Buckets

cs72                             6            0                  256           << - group 20

cs70                             6            0                  386           <<<<<<<<<<<<<<<<<<<<<- group 10 hash problem !!

cs71                             6            0                  -130          <<<<<<<<<<<<<<<<<<<<< - group 10 hash problem !!

..

In this case, I disabled and enabled CS70 traffic was distributed to CS70 instead CS71 even though I waited over 10 minutes but nothing matter even though “csw-force-rehash” is using. Until, I removed and put “default hash-persist”  back. (default hash-persist url search-url "" offset 0 length 80). Everything become alright like result below and traffic is distributed to both cache server (CS70 and CS71).

############ BP 1/1 output start #############

Cache Server Name                Admin-status Hash-distribution  L7-Hash-Buckets

cs72                             6            0                  256           << - group 20

cs70                             6            0                  129           <<<<<<<<<<<<<<<<<<<<< group 10 half and half

cs71                             6            0                  127           <<<<<<<<<<<<<<<<<<<<< group 10 half and half

In another case, I put “max-conn 1000” command into cache servers. The problem like above happened again and couldn’t work around by re-enabling “default hash-persist” anymore until reboot server. Please see testing scenarion in the picture below:

max-conn_issue_resize.jpg

Please give me your suggestion for this case

Super Contributor
Posts: 1,087
Registered: ‎12-13-2009

Re: Persistent hash problem after configuring max-conn, disable/enable

Ho Tony,

     That to me looks like a bug to me.  Suggest you log a TAC ticket for this one.

     What version of code do you have? Might be best to check the releanse note for the latest version to see if it is a kown bug or not.

Thanks

Michael.

Occasional Contributor
Posts: 17
Registered: ‎11-10-2010

Re: Persistent hash problem after configuring max-conn, disable/enable

Hi Michael,

It's using version 12.2.01T403

Thank you so much!!!

Tony

Occasional Contributor
Posts: 17
Registered: ‎11-10-2010

Re: Persistent hash problem after configuring max-conn, disable/enable

Hi Michael,

Cool!!! I upgraded software to lastest version (Version 12.2.01bT403). The release-note mentions about fixing many bugs (TCS as well). I testes with many scenarios it's working !! The result shows like example below:

2_DisableCS70WithoutClearHashBucket_resize.jpg

Cheers,

Tony

Super Contributor
Posts: 1,087
Registered: ‎12-13-2009

Re: Persistent hash problem after configuring max-conn, disable/enable

Hi Tony,

     Glad you got a fix to the issue.

Thanks

Michael.

Join the Community

Get quick and easy access to valuable resource designed to help you manage your Brocade Network.