Closed jssmith closed 6 years ago
that almost certainly means that the connection is dropped. we can add that to the retry path
Ok. What do you think about possible fragmentation? I'm not sure what sort of network layers / abstractions we might be sitting on top of in this environment. I will try wrapping the read in a loop.
I might be wrong because I haven't run without blocking turned off in forever, but I think* we should be waiting until all the bytes have been received.
unless I misunderstand your question, but by all means dig into it. one of the things I'd like to do is have a sprintf thing so we can add additional information to the status strings
I'm pretty sure I found it—we don't always read a full block. E.g., after adding logging:
read 192.168.1.57/tpcc-nfs offset:99737600 bytes:4096 partial read 2892 of 4200
On the logfix branch.
sorry, didn't think that was supposed to happen without fionbio
All good.
When running TPC-C in Lambda error messages sometimes appear that are not seen when running in EC2, e.g.:
this leads to SQLite failed transactions such as: