Open jolly-fellow opened 1 year ago
@jolly-fellow based off discussion in refinement today, it sounds like you may have had updated errors you received when troubleshooting with Kevin, can you update this ticket with these? We were also curious whether there were any power saving functions on your device active that may have impacted the contract deployment time duration?
Can you also share your host OS? Is there any chance your Docker settings may be impacting there either?
Host OS is Ubuntu 20.04.5 LTS on both CPU. Docker settings are the same and default. Docker image is the same on both CPU. OS on the image is the same Ubuntu 20.04.5 LTS. Phenom II X4 955 has no energy saving functions. The OS on Phenom CPU worked with default governor "ondemand". Unfortunately I have no new information about this error. I stopped to spend time to this error when we decided that it is practically senseless to find optimal settings for work on Phenom CPU because the users will not do so many changes. It is easier to upgrade the hardware and it may happens that for another CPU values of the timeouts will be different. Another problem I get the same error from time to time on Ryzen CPU too. Kevin said it is OK and this is the cause why the example of bootstrap process script bios-boot-tutorial.py is full of timeouts and calling of cleos repeating until successful execution.
EOS bootstrap doesn’t work at all on Phenom II X4 955 doesn’t matter where it was run in DUNE or in bios-boot-tutorial.py . Both sequences are not working correctly on Phenom CPU because of timeouts but both work excellent on Ryzen 7 5800H. I tried to force the EOS works on the Phenom but without positive result. When I was increasing one timeout the system hit to another one. More details can be found here: https://github.com/AntelopeIO/DUNE/pull/93
Actually timeout errors sometimes happens even on fast CPU. This is why bios-boot-tutorial.py retry most of EOS command calls until get successful result.