Insufficient resources

renu

We have 3 nodes with the following RAM: Node 1 - 4GB Node 2 - 2GB Node 3 - 8GB These are in a cluster. I am trying to load data onto Vertica using both COPY & INSERT commands with DIRECT (command and hint as applicable) by connecting to the database from Node 1 We are facing the error ERROR 3587: Insufficient resources to execute plan on pool general [Request Too Large:Memory(KB) Exceeded: Requested = 687627, Free = 423576 (Limit = 423576, Used = 0)] Free = 423576 is just the RAM free space on Node 1. The RAM space available in the cluster is not being considered for loading the data. Could you please suggest how this can be resolved? Appreciate your quick help and response

Prasanta_Pal

Hi, Execute the 'top' command to see what's eating the memory. and also exceute dbadmin=> select * from RESOURCE_ACQUISITIONS; to see who has acquired the resources. Also check the Platform Requirements and Recommendations: https://my.vertica.com/docs/6.1.x/HTML/index.htm#18671.htm

renu

renu

Hi Prasanta Thanks for the reply PFB output of top 7 processes Tasks: 228 total, 1 running, 226 sleeping, 1 stopped, 0 zombie Cpu(s): 1.3%us, 1.0%sy, 0.0%ni, 95.9%id, 1.8%wa, 0.0%hi, 0.0%si, 0.0%st Mem: 3877708k total, 3047996k used, 829712k free, 243180k buffers Swap: 4095992k total, 76k used, 4095916k free, 1333256k cached PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 9030 dbadmin 20 0 2846m 322m 18m S 1.3 8.5 14:09.03 vertica 3269 dbadmin 19 -1 1392m 33m 6276 S 1.0 0.9 14:49.01 python 3080 root 20 0 1403m 56m 4472 S 0.7 1.5 5:41.74 python 24820 dbadmin 20 0 15088 1324 940 R 0.7 0.0 0:00.09 top 2375 root 20 0 8404 588 480 S 0.3 0.0 0:00.78 fcoemon 3118 cloudera 20 0 3519m 593m 16m S 0.3 15.7 3:08.19 java 1 root 20 0 19404 1552 1244 S 0.0 0.0 0:00.84 init On running the query, few observations are: is_execution is false for all rows, which indicates there are no open running processes which may compete for RAM All request types so far are for Reserve. Total memory in use is 26GB (believe this is correct since we have loaded around 26GB of data using direct command) Could you please let us know if Vertica is considering the total available RAM space on all all 3 nodes of my cluster when I load data from node01 or is it considering the RAM space on node01 only? I receive this error when I try to select 1.5GB of data from the tables loaded onto Vertica. Do I need to add more RAM or more nodes to my cluster? Would this help? Regards Renu

[Deleted User]

Hi Renu (and hi Prasanta!, I hope you don't mind if I jump in briefly) It seems like you're on a good line of debugging; hopefully Prasanta can help you with it. But I would like to second Prasanta's request that you carefully re-read the Platform Requirements document. It sounds like this cluster does not meet various of Vertica's requirements (particularly about homogeneous hardware, also I bet about minimum/recommended RAM per CPU core), and is running into the sorts of problems described by that document as a result. If you've read it and have further questions, feel free to ask. Adam

vaibhav_1

ss