Exporting traces timeouts

[lukaskai]

Hello, I have an issue with exporting the traces from Parity. Setting batches even to 1 block per file, and timeout to 600 seconds, the export of certain block traces timeouts.

A very good example block: 3805005. It always timeouts and interestingly there are over 100 failed internal txs (https://etherscan.io/txsInternal?block=3805005). Therefore, it always timeouts wihle running the export script. Or same with block: 3805016.

Other example block batch also timeouts: 2370400- 2370500. There are more as well.

Anyone having simillar issues or ideas how to fix?

[medvedev1088]

Thanks for reporting this issue!

What instance type are you using for the Parity node? I've been using n1-highmem-2 (2 vCPUs, 13 GB memory) in GCP and all the blocks are exported within the default timeout. I suspect the bottleneck may be memory in your case?

[tokusyu]

I had same problem that occurrd same block range.
After I changed DEFAULT_IPC_TIMEOUT to 6000, the timeout never occurr.
My environment is AWS m5 large(2 vCPUs, 8 GB memiry) instance with EBS(type: gp2, IOPS:7500)

[lukaskai]

Thanks, it worked. It seems that I had too low IPC_Timeout set up, as well as instance setup was not right. With m5 large it seems to be working like charm. Thanks guys, appreciate you help.

[cxd1995]

@tokusyu @lukaskai
I had same problem. Could you teach me how to change default _ipc_timeout?

[logiclord]

export DEFAULT_TIMEOUT=6000 or --timeout 6000 flag should work