NimbostratusHello
can some explain what exactly "static persist" load balancing method is? and when this method is used should there any persistence applied at the WIDE IP level ?
set up here is , WIDE-IP (abc.com) - with 1 pool which is 1 DC with 2 virtual servers in the pool and pool has static persit as preferred LB method.
so when a LDNS server is making a dns query for abc.com how is it applying persistence
thanks in advance.
solmon
Nimbostratusto further add at WIDE-IP level LB is global availability
Nimbostratus
AltostratusThis should help: GTM Static Persistence.
GTM utilizes the LDNS src IP in a hash that reorders the virtual servers in the pool. GTM then chooses the first available VS from the pool.
AltostratusThe persistence settings on the wip are separate from the static persist pool lb method.
The wip persistence settings cause GTM to create a persistence record with a timeout, based on the result of the pool lb method. This allows the pool to use a dynamic lb method (least connections for example) to select the best pool member and then the ldns will persist to it based on the persistence record. On the other hand the pool static persist lb method can be used without any wip persistence settings. By itself static persist does not create any records and has no timeout. As long as the ldns src IP remains the same and the number of available vs in the pool remains the same then the result of the static persist hash will always select the same vs IP for the dns response. That being the case, your GTM is using static persist as preferred pool lb method which likely selects the same vs for every query from a given ldns and then, in addition, the wip persist creates a persistence record for the ldns>vs ip. This is redundant. Based on your dns lb goals one or the other persistence options would suffice.
Also, the wip lb methods purpose is to select a pool but since there is only one pool assigned to the wip the wip lb method is irrelevant.