-
WAL and MANIFEST might be good candidates for using direct I/O. The `LogWriter` already handles organizing writes into contiguous blocks. I'm not sure what impact, if any, on performance direct I/O wo…
-
```
It is not a bug or issue.
It is just a comment for tuning of SyncDirIfManifest().
I think fdatasync() is enough to sync directory and can show a little bit
better performance.
diff --git a/util…
-
Seems that rrdcached does not make sure that data is written to disk, not even the journal is synced to disk.
I am unable to find and calls to fsync, syncfs, fdatasync ...
So I assume that a consi…
-
Unsure if anybody needs this. I might need this because in one usecase I'm writing _a lot_ of files (to different filenames), and only need a guarantee that a SIGKILL won't leave a partially written f…
-
I found weird situation: if I constantly use `fsync` application works steadily, but if I change it on `fsyncdata`, it sometimes returns error: `Bad file descriptor (os error 9)`. In linux man I found…
-
```
It is not a bug or issue.
It is just a comment for tuning of SyncDirIfManifest().
I think fdatasync() is enough to sync directory and can show a little bit
better performance.
diff --git a/util…
-
```
It is not a bug or issue.
It is just a comment for tuning of SyncDirIfManifest().
I think fdatasync() is enough to sync directory and can show a little bit
better performance.
diff --git a/util…
-
to make sure flush_lsn initialization is correct
-
**systemd version the issue has been seen with**
> v244
**Used distribution**
> …
**Expected behaviour you didn't see**
> To see fsync after the creating, removing and cleaning oper…
-
I think this package would be more useful if there was a call to fsync() as well, at least optionally, as with http://hackage.haskell.org/package/safeio-0.0.5.0/docs/src/System-IO-SafeWrite.html#withO…