job picked up by render-node from different pool

[ultra-sonic]

Hi Timur,

too mitigate the "capacity dilemma" #540 I have written a python script that locks jobs into different pools.
i have a pool for 256 thread machines, one for 192 thread machines and so on..

the script sent this payload to job id 139

{
    "action": {
        "user_name": "xxxxx", 
        "ids": [
            139
        ], 
        "type": "jobs", 
        "host_name": "xxxxxx", 
        "params": {
            "pools": {
                "/berlin/station/32": 99, 
                "/berlin/farm/32": 99
            }
        }
    }
}

still the job is being picked up by render-nodes from pool /berlin/farm/256
any idea why this can happen?

[ultra-sonic]

do i have to enable pool solving somewhere?

[ultra-sonic]

or do i need to restart afserver after i created all the pools?

[timurhai]

Hi Oliver!

This is a priority, not a hosts mask.
/berlin/station/32 has 99 and /berlin/farm/256 has 0, but not disabled.
But, If some pool priority is 100 - zero (not specified) pools will be disabled (may say that works like hosts mask).

Pools solving is always enabled and can't be disabled.
After pools manipulation afserver restart not needed.

[timurhai]

Also, I forgot to say, that priority -100 disables pool.

[ultra-sonic]

i dont really understand.if i have these pools:

/berlin/farm/32
/berlin/farm/64
/berlin/farm/128
/berlin/farm/256

what would be the payload to render a job on /berlin/farm/32 only and disable the other 3 pools for the jobs?

[timurhai]

"/berlin/farm/32": 100

[ultra-sonic]

ok...thanks...i assume to render it on 2 pools it would be this?

"/berlin/farm/32": 100,
"/berlin/farm/64": 100

why does a priority of 99 work completly different?

[timurhai]

1. Yes.
2. To make pools to work like hosts mask within the same priority parameter, to not to add more parameter(s).