Commit Graph

210 Commits

Author SHA1 Message Date
Simon Frei
1f8e6c55f6
lib/db: Refactor to use global list by version (fixes #6372) (#6638)
Group the global list of files by version, instead of having one flat list for all devices. This removes lots of duplicate protocol.Vectors.

Co-authored-by: Jakob Borg <jakob@kastelo.net>
2020-05-30 09:50:23 +02:00
Jakob Borg
94beed5c10
lib/db: Add Badger backend (fixes #5910) (#6250) 2020-05-29 13:43:02 +02:00
Jakob Borg
f78133b8e9
lib/db: Adjust transaction flush sizes downwards (#6686)
This reduces the size of our write batches before we flush them. This
has two effects: reducing the amount of data lost if we crash when
updating the database, and reducing the amount of memory used when we do
large updates without checkpoint (e.g., deleting a folder).

I ran our SyncManyFiles benchmark as it is the one doing most
transactions, however there was no relevant change in any metric (it's
limited by our fsync I expect). This is good as any visible change would
just be a decrease in performance.

I don't have a benchmark on deleting a large folder, taking that part on
trust for now...
2020-05-27 12:15:00 +02:00
Simon Frei
8f344d0915
lib/db: Initialize need meta on first dev occurrence (fixes #6668) (#6669) 2020-05-20 11:01:27 +02:00
Simon Frei
5b34c31cb3
lib/db: Don't panic on seq. coruption when debugging (#6662) 2020-05-18 10:42:51 +02:00
Audrius Butkevicius
a1c5b44c74
lib/model: Fix rename handling (ref #6650) (#6652) 2020-05-16 14:39:27 +02:00
Simon Frei
de9489585f
lib/db: Filter repeat files in one update (ref #6650) (#6651) 2020-05-16 14:34:53 +02:00
Simon Frei
974551375e
lib/db: Dont add symlinks to blocks map (fixes #6637) (#6639) 2020-05-13 20:38:21 +02:00
Jakob Borg
531ceb2b0f
Add indirection for large version vectors. (#6376)
This adds indirection of large version vectors in the same manner as we
already to block lists. The effect is the same: less duplicated data in
some situations.

To mitigate the impact for when this indirection
wouldn't be needed I've added an indirection cutoff for both blocks and
the new version vector stuff: we don't do the indirection at all for
small block lists or small version vectors, instead storing it directly
like we used to do. This is faster for small files and small setups.
2020-05-13 14:28:42 +02:00
Simon Frei
ba4462a70b
lib/db: Fix updating/removing from global (ref #6413) (#6635) 2020-05-13 12:56:49 +02:00
Audrius Butkevicius
decb967969
all: Reorder sequences for better rename detection (#6574) 2020-05-11 20:15:11 +02:00
Simon Frei
a94951becd
lib/db, lib/model: Keep need stats in metadata (ref #5899) (#6413) 2020-05-11 15:07:06 +02:00
Audrius Butkevicius
da99203dcd
lib/db: Fix non-truncated fileinfo loading (#6621) 2020-05-08 20:31:43 +01:00
Jakob Borg
c0c18a568c
lib/db: Hold the bloom filter the right way (fixes #6614) (#6617) 2020-05-08 14:18:00 +02:00
Simon Frei
22242d51be
lib/db: Refactor getting global lists (#6529)
* lib/db: Refactor getting global lists

* VL -> Versions

* keyBuf

* don't touch db migration
2020-05-01 08:30:20 +01:00
greatroar
0e5ba3ca05
lib/db: Upgrade to Blobloom v0.1.1 (#6553)
Now faster and Apache-licensed.
2020-04-20 14:23:36 +02:00
greatroar
44b0f0b456
lib/db: Switch to faster blobloom Bloom filter pkg (#6537) 2020-04-20 09:02:33 +02:00
greatroar
82fbcb96f8
lib/db: Remove unused blockFinder global (#6534) 2020-04-14 17:09:59 +02:00
Simon Frei
0ba3abdee4
lib/db: Handle missed error variable in old schema upgrade (#6528) 2020-04-13 22:58:04 +02:00
Simon Frei
ab92f8520c
cmd/syncthing, lib/db: Store upgrade info to throttle queries (fixes #6513) (#6514) 2020-04-13 10:21:07 +02:00
Jakob Borg
0e67c036bb
lib/db: Make database GC a service, stop on Stop() (#6518)
This makes the GC runner a service that will stop fairly quickly when
told to.

As a bonus, STTRACE=app will print the service tree on the way out,
including any errors they've flagged.
2020-04-12 10:26:57 +02:00
Simon Frei
df318ed370
lib/db: Don't get blocklists on drop and missing continue (ref #6457) (#6502) 2020-04-07 13:13:18 +02:00
Simon Frei
32245435e2
lib/model: Handle deleted items on RO for remote removes (fixes #6432) (#6464) 2020-04-02 16:14:25 +02:00
greatroar
2f26a95973
lib/db, lib/model: Code simplifications (#6484)
NamespacedKV.prefixedKey is still small enough to be inlined.
2020-03-31 14:32:24 +02:00
Simon Frei
7f23de4f03
all: Pass db intervals as args not env vars (#6448) 2020-03-24 13:53:20 +01:00
Jakob Borg
c4abe6f815
lib/db: Don't whack blocks when putting truncated file (#6434)
As of the latest database checker we are again putting files without
blocks. I'm not 100% convinced that's a great idea, but we also do it
for ignored files apparently so it looks like we probably should support
it. This adds an escape hatch that must be manually enabled...
2020-03-20 12:07:14 +01:00
Simon Frei
74706bb02b
lib/db, lib/syncthing: Repair db once on upgrade (ref #6425, #6427) (#6429) 2020-03-19 15:58:32 +01:00
Simon Frei
cf11fa4327
lib/db: Fix removeFromGlobal and no filenames in error (fixes #6427) (#6428) 2020-03-19 14:32:22 +01:00
Simon Frei
40580d8b9b
lib/db: Remove emptied global list in checkGlobals (fixes #6425) (#6426) 2020-03-19 14:30:20 +01:00
Simon Frei
00b2340f9a
lib/db: Checkpoint during schema updates (fixes #6422) (#6424) 2020-03-18 20:33:43 +01:00
Simon Frei
cc2a55892f
lib: Repair sequence inconsistencies (#6367) 2020-03-18 17:34:46 +01:00
Simon Frei
1bd4ea0cbb
lib/db: Don't ignore failures unmarshaling version lists (#6411) 2020-03-16 10:09:27 +01:00
Simon Frei
a1cb1d70c4
lib/db: Use need func in withNeed and simplify (#6412) 2020-03-16 08:45:45 +01:00
Simon Frei
16698b12b1
lib/db: Extend set test with second remote (#6402) 2020-03-11 08:15:45 +01:00
Jakob Borg
2faa1ad360
lib/db: Be more lenient during migration (fixes #6397) (#6398) 2020-03-06 20:50:55 +01:00
Jakob Borg
dd92b2b8f4
all: Tweak error creation (#6391)
- In the few places where we wrap errors, use the new Go 1.13 "%w"
  construction instead of %s or %v.

- Where we create errors with constant strings, consistently use
  errors.New and not fmt.Errorf.

- Remove capitalization from errors in the few places where we had that.
2020-03-03 22:40:00 +01:00
Jakob Borg
c08e253e7c
lib/db: Prevent GC concurrently with migration (fixes #6389) (#6390) 2020-02-29 19:51:32 +01:00
Jakob Borg
5de6f6d349
lib/db: Correct metadata recalculation (fixes #6381) (#6382)
If we decide to recalculate the metadata we shouldn't start from
whatever we loaded from the database, as that data is wrong. We should
start from a clean slate.
2020-02-28 11:16:33 +01:00
Jakob Borg
883497966e lib/db: Remove reference to env var that never existed 2020-02-27 11:21:35 +01:00
Jakob Borg
4f7a77597e
lib/db: Slightly improve indirection (ref #6372) (#6373)
I was working on indirecting version vectors, and that resulted in some
refactoring and improving the existing block indirection stuff. We may
or may not end up doing the version vector indirection, but I think
these changes are reasonable anyhow and will simplify the diff
significantly if we do go there. The main points are:

- A bunch of renaming to make the indirection and GC not about "blocks"
  but about "indirection".

- Adding a cutoff so that we don't actually indirect for small block
  lists. This gets us better performance when handling small files as it
  cuts out the indirection for quite small loss in space efficiency.

- Being paranoid and always recalculating the hash on put. This costs
  some CPU, but the consequences if a buggy or malicious implementation
  silently substituted the block list by lying about the hash would be bad.
2020-02-27 11:19:21 +01:00
Jakob Borg
10cb14fcb8 lib/db: Allow put partial FileInfo without blocks (ref #6353) 2020-02-22 17:44:34 +01:00
Simon Frei
32e12abb43
lib/db: Don't panic on incorrect BlocksHash (fixes #6353) (#6355) 2020-02-22 16:51:23 +01:00
Simon Frei
6489feb1d7
lib/db: Schema update to repair sequence index (ref #6304) (#6350) 2020-02-22 09:36:59 +01:00
Simon Frei
fae7425bbf
lib: Modify FileInfos consistently (ref #6321) (#6349) 2020-02-19 16:58:09 +01:00
Simon Frei
05e23f1991
lib/db: Don't call backend.Commit twice (ref #6337) (#6342) 2020-02-14 08:11:24 +01:00
Jakob Borg
6a840a040b
lib/db: Keep metadata better in sync (ref #6335) (#6337)
This adds metadata updates to the same write batch as the underlying
file change. The odds of a metadata update going missing is greatly
reduced.

Bonus change: actually commit the transaction in recalcMeta.
2020-02-13 15:23:08 +01:00
Simon Frei
ca90f4e6af
lib/db: Use flags from arg not LocalFlags() updating meta (#6333) 2020-02-13 14:02:30 +01:00
Jakob Borg
51fa36d61f
lib/db: Recover sequence number and metadata on startup (fixes #6335) (#6336)
lib/db: Recover sequence number and metadata on startup (fixes #6335)

If we crashed after writing new file entries but before updating
metadata in the database the sequence number and metadata will be wrong.
This fixes that.
2020-02-13 13:05:26 +01:00
Jakob Borg
d95a087829
lib/db: Don't leak snapshot when closing (#6331)
We could potentially get a snapshot and then fail to get a releaser,
leaking the snapshot. This takes the releaser first and makes sure to
release it on snapshot error.
2020-02-12 12:00:17 +01:00
Jakob Borg
a728743c86
lib/db: Use Commit() instead of commit() (#6330)
The readWriteTransaction offered both commit() (the one to use) and
Commit() (via embedding) where the latter didn't close the read
transaction. This removes the lower cased variant in order to prevent
the mistake.

The only place where the mistake was made was the new gc runner, where
it would leave a read snapshot open forever.
2020-02-12 11:59:55 +01:00