Daniel Farina

Firstly, cruft itself is a problem. It should not build up unless WAL-E is crashing all the time in a way it cannot clean up, which is strange. As for...

@drob yes. Or, setting things up in such a way that the tempfile is unlinked immediately (so when WAL-E exits it'll be reclaimed by the file system) may be a...

Another matter to take into account with multi-part is that it's a whole different kind of object in the S3 Parthenon: half uploads would need a way to be cleaned...

It might be that it's possible to marry multi-part and the 'delete' reaping code without too much mechanism, in which case neat. On the other hand, the multi-key approach is...

On Wed, Apr 13, 2016 at 6:48 PM Eugene Trufanov [email protected] wrote: > Hi, > It looks like this issue isn't resolved, and TMPDIR workaround isn't > working as well...

So, interestingly, I thought Near The Beginning Years Ago that this feature would be necessary for WAL-E, since then I have encountered WAL-gaps literally zero times, in which time WAL-E...

Basically submit something that touches the features you want but don't dwell too much on testing or naming things or refactoring. I would estimate such a prototype might take 25%...

There are a couple of ways to do this, probably more roundabout than you'd like but also compatible with more archivers. One is to monitor the postgres `archive_status` directory, e.g....

The problem is I have a gevent dependency, and there is a Python3 variant but it's still in beta. I have done some dry-runs of taking WAL-E to Python3 and...

Oh I see, boto3 is not python3 boto, it's just newer boto. Sure, why not, I guess.