Daniel Farina
Daniel Farina
Are you hosting in Azure? I am wondering if your link to WABS was slow enough somehow that it caused timeouts.
I don't have any specific advice on the matter, no.
Oh, yes, that would be good. I will cut a release soon.
What storage backend is that? Are you using the `WALE_S3_ENDPOINT` env var?
Yes, thank you, the error message should be better. I'm not sure if I'll fix it for now (will you?) but let's leave this open.
Yeah, we've had this happen too. @gregburek suffered mightily for it; our solution is to use `TMPDIR` set to a tmpfile-ish directory, using the ephemeral disk on many Amazon instances....
Well, again, `TMPDIR`, which is a standard unix-ism, but one I suppose can be informed that a given program actually cares about this norm. That perhaps can go into the...
Also, this discussion in burying the lede: the error message that occurs when this happens is not very good. `ENOSPC` should probably emit a hint message.
Also, the problem of using `TMPDIR` can be fixed in its entirety by using multi-part upload and a bit of buffering, see https://github.com/kr/s3/tree/master/s3cp. I (or anyone else) just haven't gotten...
The base backup push can be pretty stressful on the system (optimizations/profiling welcome), consider the rate-limiting feature. I use 10MB/s and that seems to work pretty well.