Note: I am running code in a cluster with 16 slaves, HPCC version 6.4.40
I am running some ECL code that returns this error:
System error: 0: Graph graph2[14], SLAVE #1 [10.313.316.31:20100]: Error receiving actinit data for graph: 14
What does this error exactly indicates?
Is it maybe running out of memory?
In the thor master log just before the exception I can see there are two lines of log, first one starting by NIC (Network Interface?) and other with SYST (System?)
Values doesn't seem to change drastically:
From the development team:
Why you are seeing that error:
There are a lot of logical files at the same scope level, causing significant access (lookup) slowdowns, ultimately meaning that if there are 100's or 1000's being looked up for a single read, it is exceeding the timeout.
Scope's with a lot of logical files at the same level like this used to be a pain point for Dali and clients accessing files within them. Basically, it caused each lookup to perform a linear search through the scope for match. NB: that was fixed some years ago (in 7.12.0)
So my guess is that the # of files in scopes being accessed by this query (that haven't been rolled up?) have grown and are now causing the cumulative time to look them up to exceed the [25 minute] timeout.
Recommend you rollup your files and/or upgrade your cluster as soon as possible. The current gold release is now up to Version 9.
Hope this helps,
Bob