Agata Kruszona-Zawadzka

Hi @deltabweb Thank you for this, it's a perfectly reasonable request. We're gonna add some logs pertaining to config files and problems with them.

And a question - since you mentioned in the other issue that you use nullfs over moosefs, is it the case here also? Or are you testing directly on mfs...

Currently there is no possibility of setting an offset to saving metadata and the server starts counting at midnight GMT, so if you set freq to 24 hours, it will...

In general, we don't want to be too greedy. Read ahead of 16 chunks, when the real read request is for, let's say, 200 bytes, makes no sense. The client...

Although, as usual, we could add an option ;) for single-point use systems that want and need to exhaust resources to be efficient.

One important question before we go any further: what is the version of the chunk server you've pulled your hdd from and what is the version of the chunk server...

Okay, so we have several separate issues here: 1) _Scanning is slow_ - scanning with duplicates is always slower. It can be slowed down even more if for some reason...

As for 5 - this is a bug in reporting. The time is 168 hours (1 week), but because of a bug it's reported backwards (it should say how much...

Well, this most definitely is not a correct chunk header. So it's not a bug in the chunk server's behaviour, it's the real corruption of chunks. The question is, why...

Thank you for all the info. We are running tests right now, we will let you know as soon as we find something.