Closed ganning127 closed 9 months ago
Consider squashing now so no one is confused by the didn't mean to delete code
commit.
squashed!
the small timeout issue right now, is that if you run docker restart nfsrods
, and then run a command like ls
immediately after (like without giving the NFSRODS instance a second or so), NFSRODS will output a short error message, before coming alive again.
$ docker restart nfsrods
nfsrods
$ ls
nfs server localhost:/: not responding
nfs server localhost:/: is alive again
anotherFolder dave.txt lightsssss.txt
asdfasdfasd.txt daveFolder macBookPro.txt
aslkdfjalksjfasdjfakasfljkasdfjlkadf.txt davvvvvveee.txt newBrah.txt
bar.txt eagle.txt newFileGanning.txt
booobooo.tasdfas foo.txt stuffs
coocoo.txt grrr.txt testFolder
daaaaa.txt laksdfljkas.txt
however, viewing, listing, and editing files all work as normal now after doing restart without re-mounting
oh... this is very cool.
that error message is the 'client side' printing that out - and it's due to the NFSRODS server not responding fast enough. not sure there's anything we can do about that except eventually work to optimize startup time (a bit out of our control as we're just providing the virtual filesystem backend to the NFS4J server framework).
the next pass through the code, please remove all the deadcode and consider some comments for 'why' you're adding the things you've added. please send flare for when you want eyeballs/comments on the code itself.
good stuff.
looking very good - need some @korydraughn eyeballs before squashing.
Will look as soon as I can.
Just leaving this here for reference. We're having some inconsistencies with some tests failing on Linux but not on macOS. Specific details are in this new issue: #187
changes made
getAndIncrementFileId()
, so we are no longer using an incremental counter as the inode numbers'inodeToPath_
andpathToInode_
hashmaps will be cleared out. thus,getPath()
will returnnull
. now, ifgetPath()
returns null, we will ask GenQuery for where we are within the filesystem