One of the causes of "panic: database is closed" is that we try to send
summaries after it's been closed. Calculating summaries can take a long
time and if we have a lot of folders it's not unreasonable to think
that we might be stopped in this loop, so prepare to bail here.
* push
During some other work I discovered these tests weren't great, so I've
rewritten them to be a little better. The real changes here are:
- Don't play games with not starting the folder and such, and don't
construct a fake folder instance -- just use the one the model has. The
folder starts and scans but the folder contents are empty at this point
so that's fine.
- Use a fakefs instead of a temp dir.
- To support the above, implement a fakefs option `?content=true` to
make the fakefs actually retain written content. Use sparingly,
obviously, but it means the fakefs can usually be used instead of an
on disk real directory.
This adds a new config with the simple and concise name
maxConcurrentIncomingRequestKiB. This limits how many bytes we have "in
the air" in the form of response data being read and processed.
After some testing I think that not having this limiter is seldom a
great idea and thus I propose a default value of 256 MiB for this new
setting.
I also refactored the folder IO limiter to be a model/folder attribute
instead of a package global.
Adds a new folder state "Waiting to Sync" in the same vein as the
existing "Waiting to Scan". This vastly improves performances in the
rare cases when there are lots and lots of folders operating.
* lib/db: Deduplicate block lists in database (fixes#5898)
This moves the block list in the database out from being just a field on
the FileInfo to being an object of its own. When putting a FileInfo we
marshal the block list separately and store it keyed by the sha256 of
the marshalled block list. When getting, if we are not doing a
"truncated" get, we do an extra read and unmarshal for the block list.
Old block lists are cleared out by a periodic GC sweep. The alternative
would be to use refcounting, but:
- There is a larger risk of getting that wrong and either dropping a
block list in error or keeping them around forever.
- It's tricky with our current database, as we don't have dirty reads.
This means that if we update two FileInfos with identical block lists in
the same transaction we can't just do read/modify/write for the ref
counters as we wouldn't see our own first update. See above about
tracking this and risks about getting it wrong.
GC uses a bloom filter for keys to avoid heavy RAM usage. GC can't run
concurrently with FileInfo updates so there is a new lock around those
operation at the lowlevel.
The end result is a much more compact database, especially for setups
with many peers where files get duplicated many times.
This is per-key-class stats for a large database I'm currently working
with, under the current schema:
```
0x00: 9138161 items, 870876 KB keys + 7397482 KB data, 95 B + 809 B avg, 1637651 B max
0x01: 185656 items, 10388 KB keys + 1790909 KB data, 55 B + 9646 B avg, 924525 B max
0x02: 916890 items, 84795 KB keys + 3667 KB data, 92 B + 4 B avg, 192 B max
0x03: 384 items, 27 KB keys + 5 KB data, 72 B + 15 B avg, 87 B max
0x04: 1109 items, 17 KB keys + 17 KB data, 15 B + 15 B avg, 69 B max
0x06: 383 items, 3 KB keys + 0 KB data, 9 B + 2 B avg, 18 B max
0x07: 510 items, 4 KB keys + 12 KB data, 9 B + 24 B avg, 41 B max
0x08: 1349 items, 12 KB keys + 10 KB data, 9 B + 8 B avg, 17 B max
0x09: 194 items, 0 KB keys + 123 KB data, 5 B + 634 B avg, 11484 B max
0x0a: 3 items, 0 KB keys + 0 KB data, 14 B + 7 B avg, 30 B max
0x0b: 181836 items, 2363 KB keys + 10694 KB data, 13 B + 58 B avg, 173 B max
Total 10426475 items, 968490 KB keys + 9202925 KB data.
```
Note 7.4 GB of data in class 00, total size 9.2 GB. After running the
migration we get this instead:
```
0x00: 9138161 items, 870876 KB keys + 2611392 KB data, 95 B + 285 B avg, 4788 B max
0x01: 185656 items, 10388 KB keys + 1790909 KB data, 55 B + 9646 B avg, 924525 B max
0x02: 916890 items, 84795 KB keys + 3667 KB data, 92 B + 4 B avg, 192 B max
0x03: 384 items, 27 KB keys + 5 KB data, 72 B + 15 B avg, 87 B max
0x04: 1109 items, 17 KB keys + 17 KB data, 15 B + 15 B avg, 69 B max
0x06: 383 items, 3 KB keys + 0 KB data, 9 B + 2 B avg, 18 B max
0x07: 510 items, 4 KB keys + 12 KB data, 9 B + 24 B avg, 41 B max
0x09: 194 items, 0 KB keys + 123 KB data, 5 B + 634 B avg, 11484 B max
0x0a: 3 items, 0 KB keys + 0 KB data, 14 B + 17 B avg, 51 B max
0x0b: 181836 items, 2363 KB keys + 10694 KB data, 13 B + 58 B avg, 173 B max
0x0d: 44282 items, 1461 KB keys + 61081 KB data, 33 B + 1379 B avg, 1637399 B max
Total 10469408 items, 969939 KB keys + 4477905 KB data.
```
Class 00 is now down to 2.6 GB, with just 61 MB added in class 0d.
There will be some additional reads in some cases which theoretically
hurts performance, but this will be more than compensated for by smaller
writes and better compaction.
On my own home setup which just has three devices and a handful of
folders the difference is smaller in absolute numbers of course, but
still less than half the old size:
```
0x00: 297122 items, 20894 KB keys + 306860 KB data, 70 B + 1032 B avg, 103237 B max
0x01: 115299 items, 7738 KB keys + 17542 KB data, 67 B + 152 B avg, 419 B max
0x02: 1430537 items, 121223 KB keys + 5722 KB data, 84 B + 4 B avg, 253 B max
...
Total 1947412 items, 151268 KB keys + 337485 KB data.
```
to:
```
0x00: 297122 items, 20894 KB keys + 37038 KB data, 70 B + 124 B avg, 520 B max
0x01: 115299 items, 7738 KB keys + 17542 KB data, 67 B + 152 B avg, 419 B max
0x02: 1430537 items, 121223 KB keys + 5722 KB data, 84 B + 4 B avg, 253 B max
...
0x0d: 18041 items, 595 KB keys + 71964 KB data, 33 B + 3988 B avg, 101109 B max
Total 1965447 items, 151863 KB keys + 139628 KB data.
```
* wip
* wip
* wip
* wip
Makes the logic a bit clearer and safer. This also sneakily redefines
the 0 interval to also mean disabled, whereas it previously meant ...
sometimes default to 1s, sometimes just spin.
This adds error returns to model methods called by the protocol layer.
Returning an error will cause the connection to be torn down as the
message couldn't be handled. Using this to signal that a folder isn't
currently available will then cause a reconnection a few moments later,
when it'll hopefully work better.
Tested manually by running with STRECHECKDBEVERY=0 on a nontrivially
sized setup. This panics reliably before this patch, but just causes a
disconnect/reconnect now.
As foretold by the prophecy, "once the database refactor is merged, then
shall appear a request to propagate errors from the store known
throughout the land as the NamedspacedKV, and it shall be good".
This PR does two things, because one lead to the other:
- Move the leveldb specific stuff into a small "backend" package that
defines a backend interface and the leveldb implementation. This allows,
potentially, in the future, switching the db implementation so another
KV store should we wish to do so.
- Add proper error handling all along the way. The db and backend
packages are now errcheck clean. However, I drew the line at modifying
the FileSet API in order to keep this manageable and not continue
refactoring all of the rest of Syncthing. As such, the FileSet methods
still panic on database errors, except for the "database is closed"
error which is instead handled by silently returning as quickly as
possible, with the assumption that we're anyway "on the way out".
* lib/versioner: Reduce surface area
This is a refactor while I was anyway rooting around in the versioner.
Instead of exporting every possible implementation and the factory and
letting the caller do whatever, this now encapsulates all that and
exposes a New() that takes a config.VersioningConfiguration.
Given that and that we don't know (from the outside) how a versioner
works or what state it keeps, we now just construct it once per folder
and keep it around. Previously it was recreated for each restore
request.
* unparam
* wip
Assume a folder error was set due to bad ignores on the latest scan.
Previously, doing a manual rescan would result in:
1. Clearing the folder error, which schedules (immediately) an fs
watcher restart
2. Attempting to load the ignores, which fails, so we set a folder
error and bail.
3. Now the fs watcher restarts, as scheduled, so we trigger a scan.
Goto 1.
This change fixes this by not clearing the error until the error is
actually cleared, that is, if both the health check and ignore loading
succeeds.
* lib/model: Don't panic on failed chmod-back on directory (fixes#5836)
This makes the "in writable dir"-wrapper log chmod-back errors instead
of panicking. To do that we need a logger so the function moved into the
model package which is also the only place it's used. The tests came
along.
(The test also exercised osutil.RenameOrCopy like some sort of
piggybacking. I removed that.)
* lib/fs, lib/model: Add error channel to Watch to avoid panics (fixes#5697)
* forgot unsupported watch
* and more non(-standard)-unixy fixes
* and windows test
* review
The check in ClusterConfig() when iterating through announced devices
in a folder explicitly skips entries without a non-zero IndexID.
Therefore, the check for IndexID == 0 just below will never be true
and the intended cleanup of local index data will not happen.
Plainly remove that check to make the intended case distinction work.
* lib/protocol: Wait for reader/writer loops on close (fixes#4170)
* waitgroup
* lib/model: Don't hold lock while closing connection
* fix comments
* review (lock once, func argument) and naming
* cmd/syncthing, lib/gui: Separate gui into own package (ref #4085)
* fix tests
* Don't use main as interface name (make old go happy)
* gui->api
* don't leak state via locations and use in-tree config
* let api (un-)subscribe to config
* interface naming and exporting
* lib/ur
* fix tests and lib/foldersummary
* shorter URVersion and ur debug fix
* review
* model.JsonCompletion(FolderCompletion) -> FolderCompletion.Map()
* rename debug facility https -> api
* folder summaries in model
* disassociate unrelated constants
* fix merge fail
* missing id assignement
* lib/tlsutil: Enable TLS 1.3 when available, on test builds (fixes#5065)
This enables TLS 1.3 negotiation on Go 1.12 by setting the GODEBUG
variable. For now, this just gets enabled on test versions (those with a
dash in the version number).
Users wishing to enable this on production builds can set GODEBUG
manually.
The string representation of connections now includes the TLS version
and cipher suite. This becomes part of the log output on connections.
That is, when talking to an old client:
Established secure connection .../TLS1.2-TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256
and now potentially:
Established secure connection .../TLS1.3-TLS_AES_128_GCM_SHA256
(The cipher suite was there previously in the log output, but not the
TLS version.)
I also added this info as a new Crypto() method on the connection, and
propagate this out to the API and GUI, where it can be seen in the
connection address hover (although with bad word wrapping sometimes).
* wip
* wip
* cleanup Fatal in lib/config/config.go
* cleanup Fatal in lib/config/folderconfiguration.go
* cleanup Fatal in lib/model/model.go
* cleanup Fatal in cmd/syncthing/monitor.go
* cleanup Fatal in cmd/syncthing/main.go
* cleanup Fatal in lib/api
* remove Fatal methods from logger
* lowercase in errors.Wrap
* one less channel
I'm working through linter complaints, these are some fixes. Broad
categories:
1) Ignore errors where we can ignore errors: add "_ = ..." construct.
you can argue that this is annoying noise, but apart from silencing the
linter it *does* serve the purpose of highlighting that an error is
being ignored. I think this is OK, because the linter highlighted some
error cases I wasn't aware of (starting CPU profiles, for example).
2) Untyped constants where we though we had set the type.
3) A real bug where we ineffectually assigned to a shadowed err.
4) Some dead code removed.
There'll be more of these, because not all packages are fixed, but the
diff was already large enough.
This adds a folder option "CopyOwnershipFromParent" which, when set,
makes Syncthing attempt to retain the owner/group information when
syncing files. Specifically, at the finisher stage we look at the parent
dir to get owner/group and then attempt a Lchown call on the temp file.
For this to succeed Syncthing must be running with the appropriate
permissions. On Linux this is CAP_FOWNER, which can be granted by the
service manager on startup or set on the binary in the filesystem. Other
operating systems do other things, but often it's not required to run as
full "root". On Windows this patch does nothing - ownership works
differently there and is generally less of a deal, as permissions are
inherited as ACLs anyway.
There are unit tests on the Lchown functionality, which requires the
above permissions to run. There is also a unit test on the folder which
uses the fake filesystem and hence does not need special permissions.
* lib/fs, lib/model: Improve filesystem operations during tests (fixes#5422)
Introduces MustFilesystem that panics on errors and should be used for operations
during testing which must never fail.
Create temporary directories outside of testdata.
* don't do a filesystem, just a wrapper around os for testing
* fix copyright
This avoids waiting until next ping and timeout until the connection is actually
closed both by notifying the peer of the disconnect and by immediately closing
the local end of the connection after that. As a nice side effect, info level
logging about dropped connections now have the actual reason in it, not a generic
timeout error which looks like a real problem with the connection.
In a recent change (#5201) this return disappeared. The effect is that
we first shortcut the file and then also treat it normally. This results
in to database updates after each other, which are bound to end up in
the same batch. This means we remove one sequence entry and add two.
Not marking the issues as fixed, because I need to do more testing and
there are other discrepancies...
This adds a thin type that holds the state associated with the
leveldb.DB, leaving the huge Instance type more or less stateless. Also
moves some keying stuff into the DB package so that other packages need
not know the keying specifics.
(This does not, yet, fix the cmd/stindex program, in order to keep the
diff size down. Hence the keying constants are still exported.)
* lib/model, cmd/syncthing: Wait for folder restarts to complete (fixes#5233)
This is the somewhat ugly - but on the other hand clear - fix for what
is really a somewhat thorny issue. To avoid zombie folder runners a new
mutex is introduced that protects the RestartFolder operation. I hate
adding more mutexes but the alternatives I can think of are worse.
The other part of it is that the POST /rest/system/config operation now
waits for the config commit to complete. The point of this is that until
the commit has completed we should not accept another config commit. If
we did, we could end up with two separate RestartFolders queued in the
background. While they are both correct, and will run without
interfering with each other, we can't guarantee the order in which they
will run. Thus it could happen that the newer config got committed
first, and the older config commited after that, leaving us with the
wrong config running.
* test
* wip
* hax
* hax
* unflake test
* per folder mutexes
* paranoia
* race
The problem here is that we would update the sequence index before
updating the FileInfos, which would result in a high sequence number
pointing to a low-sequence FileInfo. The index sender would pick up the
high sequence number, send the old file, and think everything was good.
On the receiving side the old file is a no-op and ignored. The file
remains out of sync until another update for it happens.
This fixes that by correcting the order of operations in the database
update: first we remove old sequence index entries, then we update the
FileInfos (which now don't have anything pointing to them) and then we
add the sequence indexes (which the index sender can see).
The other option is to add "proper" transactions where required at the
database layer. I actually have a branch for that, but it's literally
thousands of lines of diff and I'm putting that off for another day as
this solves the problem...
The problem here is that we would update the sequence index before
updating the FileInfos, which would result in a high sequence number
pointing to a low-sequence FileInfo. The index sender would pick up the
high sequence number, send the old file, and think everything was good.
On the receiving side the old file is a no-op and ignored. The file
remains out of sync until another update for it happens.
This fixes that by correcting the order of operations in the database
update: first we remove old sequence index entries, then we update the
FileInfos (which now don't have anything pointing to them) and then we
add the sequence indexes (which the index sender can see).
The other option is to add "proper" transactions where required at the
database layer. I actually have a branch for that, but it's literally
thousands of lines of diff and I'm putting that off for another day as
this solves the problem...
This removes the out of disk space check from CheckHealth. The disk space is now
only checked if there are files to pull, in which case pulling those files is
stopped, but everything else (dirs, links, deletes) keeps running -> can recover
disk space through pulling.
Adds a receive only folder type that does not send changes, and where the user can optionally revert local changes. Also changes some of the icons to make the three folder types distinguishable.
This is an improvement of PR #4493 and related to (and maybe fixing) #4961
and #4475. Maybe fixing, because there is no clear reproducer for that
problem.
The previous PR added a mechanism to resurrect missing parent directories,
if there is a valid child file to be pulled. The same mechanism does not
exist for dirs and symlinks, even though a missing parent can happen for
those items as well. Therefore this PR extends the resurrection to all types
of pulled items.
In addition I moved the IsDeleted branch while iterating over
processDirectly to the existing IsDeleted branch in the WithNeed iteration.
This saves one pointless assignment and IsDeleted query. Also
We have the invalid bit to indicate that a file isn't good. That's enough for remote devices. For ourselves, it would be good to know sometimes why the file isn't good - because it's an unsupported type, because it matches an ignore pattern, or because we detected the data is bad and we need to rescan it.
Or, and this is the main future reason for the PR, because it's a change detected on a receive only device. We will want something like the invalid flag for those changes, but marking them as invalid today means the scanner will rehash them. Hence something more fine grained is required.
This introduces a LocalFlags fields to the FileInfo where we can stash things that we care about locally. For example,
FlagLocalUnsupported = 1 << 0 // The kind is unsupported, e.g. symlinks on Windows
FlagLocalIgnored = 1 << 1 // Matches local ignore patterns
FlagLocalMustRescan = 1 << 2 // Doesn't match content on disk, must be rechecked fully
The LocalFlags fields isn't sent over the wire; instead the Invalid attribute is calculated based on the flags at index sending time. It's on the FileInfo anyway because that's what we serialize to database etc.
The actual Invalid flag should after this just be considered when building the global state and figuring out availability for remote devices. It is not used for local file index entries.
To optimize WithNeed, which is called for the local device whenever an index
update is received. No tracking for remote devices to conserve db space, as
WithNeed is only queried for completion.
I'm trying to slowly clean this up a bit, and moving functionality out
into the folder types and having those methods not reach into model is
part of it. That can mean takign some odd arguments in the meantime,
some of those should probably become interfaces or properties on folder
in the long term.
The functionality was anyway mostly implemented there and not isolated
in the folderScanner type. The attempt to refactor it out in the other
direction wouldn't work given that the event loop and stuff is on
`folder`.
To newer names better reflecting their types and yet sorting together
with folder.go. Doing it now without asking because there are no open
PRs that will get killed by it, and to avoid bikeshedding the names.
The actual pull method (which is really the only thing that differs
between them) is now an interface member which gets overridden by the
subclass.
"Subclass?!" Well, this is dynamic dispatch with overriding, I guess.
Instead of walking and unmarshalling the entire db and sorting the resulting
file infos by sequence, add store device keys by sequence number in the
database. Thus only the required file infos need be unmarshalled and are already
sorted by index.
Bumping the limit to 2 * the max block size (16 MiB) is a slight
increase compared to previously. Nonetheless I think it's good to allow
us to queue one request and have one on the way in, or conversely have
one large block on the way in and be able to ask for smaller blocks from
others at the same time.
