search

schneems / puma_auto_tune

Puma performance without all the (T)pain
227 stars 19 forks source link

Doesn't seem to like running out of memory with a single worker #2

Closed markpundsack closed 10 years ago

markpundsack commented 10 years ago

When working locally, I set memory to 128MB, and the worker takes up 127MB before any traffic hits it, but as soon as traffic comes in, PumaAutoTune (understandably) freaks out.

14:40:30 web.1    | PumaAutoTune (73s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=226.58984375 puma.current_cluster_size=1
14:40:30 web.1    | PumaAutoTune (74s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=226.6015625 puma.current_cluster_size=1
14:40:30 web.1    | PumaAutoTune (74s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=226.6171875 puma.current_cluster_size=1
14:40:30 web.1    | PumaAutoTune (74s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=226.62109375 puma.current_cluster_size=1
14:40:30 web.1    | PumaAutoTune (74s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=226.6328125 puma.current_cluster_size=1
14:40:30 web.1    | PumaAutoTune (74s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=226.6484375 puma.current_cluster_size=1
14:40:30 web.1    | PumaAutoTune (74s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=226.66015625 puma.current_cluster_size=1
14:40:31 web.1    | PumaAutoTune (74s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=226.671875 puma.current_cluster_size=1
14:40:31 web.1    | PumaAutoTune (74s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=226.6875 puma.current_cluster_size=1
14:40:31 web.1    | PumaAutoTune (74s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=226.71875 puma.current_cluster_size=1
14:40:31 web.1    | PumaAutoTune (74s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=226.7578125 puma.current_cluster_size=1
14:40:31 web.1    | PumaAutoTune (75s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=226.76953125 puma.current_cluster_size=1
14:40:31 web.1    | PumaAutoTune (75s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=226.8125 puma.current_cluster_size=1
14:40:31 web.1    | PumaAutoTune (75s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=226.84765625 puma.current_cluster_size=1
14:40:31 web.1    | PumaAutoTune (75s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=226.86328125 puma.current_cluster_size=1
14:40:31 web.1    | PumaAutoTune (75s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=226.87109375 puma.current_cluster_size=1
14:40:31 web.1    | PumaAutoTune (75s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=226.88671875 puma.current_cluster_size=1
14:40:31 web.1    | PumaAutoTune (75s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=226.90234375 puma.current_cluster_size=1
14:40:32 web.1    | PumaAutoTune (75s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=226.9140625 puma.current_cluster_size=1
14:40:32 web.1    | PumaAutoTune (75s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=226.9296875 puma.current_cluster_size=1
schneems commented 10 years ago

Sorry for the delay, i wasn't "watching" this repo in github (sigh), it's fixed now. Great comment, we should add logic to not try to go lower than 2 workers.

markpundsack commented 10 years ago

NP. You mean lower than 1 worker, right?