This is a breaking change to users who might be using Kopia as a library.
### Log Format
```json
{"t":"<timestamp-rfc-3389-microseconds>", "span:T1":"V1", "span:T2":"V2", "n":"<source>", "m":"<message>", /*parameters*/}
```
Where each record is associated with one or more spans that describe its scope:
* `"span:client": "<hash-of-username@hostname>"`
* `"span:repo": "<random>"` - random identifier of a repository connection (from `repo.Open`)
* `"span:maintenance": "<random>"` - random identifier of a maintenance session
* `"span:upload": "<hash-of-username@host:/path>"` - uniquely identifies upload session of a given directory
* `"span:checkpoint": "<random>"` - encapsulates each checkpoint operation during Upload
* `"span:server-session": "<random>"` -single client connection to the server
* `"span:flush": "<random>"` - encapsulates each Flush session
* `"span:maintenance": "<random>"` - encapsulates each maintenance operation
* `"span:loadIndex" : "<random>"` - encapsulates index loading operation
* `"span:emr" : "<random>"` - encapsulates epoch manager refresh
* `"span:writePack": "<pack-blob-ID>"` - encapsulates pack blob preparation and writing
(plus additional minor spans for various phases of the maintenance).
Notable points:
- Used internal zero allocation JSON writer for reduced memory usage.
- renamed `--disable-internal-log` to `--disable-repository-log` (controls saving blobs to repository)
- added `--disable-content-log` (controls writing of `content-log` files)
- all storage operations are also logged in a structural way and associated with the corresponding spans.
- all content IDs are logged in a truncated format (since first N bytes that are usually enough to be unique) to improve compressibility of logs (blob IDs are frequently repeated but content IDs usually appear just once).
This format should make it possible to recreate the journey of any single content throughout pack blobs, indexes and compaction events.
Track and report errors separately according to the type of error:
- missing pack
- truncated pack
- unreadable content.
Add a counter stat for the contents that are
read and fully verified (via `GetContent`).
Count errors grouped by pack ID using a `CountersMap`.
This allows determining the number of referenced contents
that were missing in a particular pack.
Report the counter stats via structured logging.
---
Sample output:
$ kopia content verify --progress-interval=0.5s --download-percent=100
Listing blobs...
Listed 102 blobs.
Verifying contents...
Verified 1 contents, 0 errors, estimating...
Verified 279 contents, 0 errors, estimating...
Verified 512 of 624 contents (82.1%), 0 errors, remaining 0s, ETA 2025-09-17 23:03:38 PDT
Finished verifying contents
verifyCounters: {"verifiedContents":624,"totalErrorCount":0,"contentsInMissingPacks":0,\
"contentsInTruncatedPacks":0,"unreadableContents":0,"readContents":624,\
"missingPacks":0,"truncatedPacks":0,"corruptedPacks":0}
- use `slices.Clone`
- remove stale `.gometalinter.json`
- unexport `maintenance.dropDeletedContents`
- rename `fetchIndexBlob`
- use `require` in `TestTimeFuncWiring`
Functional changes:
- Key derivation no longer panics, instead it returns an
error on failed condition checks.
- Migration to crypto/hkdf required changes in various
function signatures, adding an error return value for
propagating key derivation errors up the calling chain.
Tests were refactored to account for the signature change as well.
Non-functional changes:
- migrate to crypto/pbkdf2
- migrate to crypto/sha3
- migrate to crypto/hkdf
- add deriveHMACSecret function to simplify caller function
- simplify deriveHMACSecret
- fix test: specify non-empty HMACSecret
* remove unnecessary type argument
* modernize with max
* unexport getPartial and update comment
* unexport getFull
* verifyNotCached helper
* use require helpers
* leverage verify[Not]Cached
* use windowsOSName const
* fix comment wrapping
* require in stat_test
* use 512 as the write size and log allocated size
* rename const to expectedMinAllocSize
* write a single byte to test file
* add TestGetBlockSizeFromCurrentFS
* require Positive
* log before invariant check
While sorting by modification time is possible with OS tools,
such as `ls`, including the directory creation time in the
name facilitates cleanup in local dev environments where
tests fail often.
- enable `forcetypeassert` linter in non-test files
- add `//nolint` annotations
- add `testutil.EnsureType` helper for type assertions
- enable `forcetypeassert` linter in test files
- nit: rename var to packCountByPrefix
- leverage impossible package
- use maps.Clone
- unexport indirectObjectID
- unexport compressed
- rename function to flushBufferLocked
- add checklocks annotations to functions that must be called under w.mu
* fix(server): fixed server-based notifications
Used TypedEventArgs instead of `any` to ensure all notification data
carries type information, allowing the server to property deserialize it.
* fix
Minor cleanups in diff tests: helper to generate (direct) object ids.
- nit: rename variables for consistency
- nit: Use `k` prefix in oids where appropriate
- nit: remove unnecessary intermediate vars
- nit: specify known map length
Adds a `--stats-only` flag to the `diff` command to reduce
output verbosity.
When `kopia diff` is invoked with the `--stats-only` flag it
only outputs the aggregate statistics of changes.
As jsonencoding is used to pass some config parameters, move it to a
package that callers external to kopia can name. This ensures they can
actually configure those parameters if needed.
* use `os.CreateTemp` in `tempfile.CreateAutoDelete`
* refactor `TestTempFile`: add `VerifyTempfile` test helper
* add test for `createUnixFallback`
* rename function to `tempfile.CreateAutoDelete`
* remove the `dir` parameter to `tempfile.CreateAutoDelete`,
only an empty string was passed, apart from test cases.
* guard against panic in TestShadowCopy
* use uint8 for clarity
* unexport writeContentAsyncAndVerify
* fix typo in test function name
* remove commented interface functions
* use atomic.Int32
* cleanups in socket server activation test
* leverage stdlib's maps and slices packages
replace uses of `golang.org/x/exp/maps`
* nit: leverage `maps.Values`
- remove spurious log message, duplicates what is being sent to the output;
- nit: update struct comment;
- consolidate code via a parameterized function;
- rename variable for clarity, it makes it easier to understand by avoiding negative condition check;
- remove unused return value;
- replace if blocks with switch.
Extends error message with a generic, OS-independent hint about
what the source of the error may be and a potential solution.
Also, modifies error message to avoid duplicate messages.
- kopia/kopia#4449
Added functionality to calculate aggregate statistics when
comparing what's changed between snapshots using kopia diff
Statistics collected during snapshot diff computation includes:
- files added/removed/modified
- dirs added/removed/modified
- files/dirs with metadata changes but same underlying content (OID)
Testing approach:
Added a test for verifying stats collected when comparing two directories with the same objectID but metadata changes across snapshots (dir mode, dir mod time, dir owner, etc), expectation is all the appropriate dir stats fields are updated.
Added another test for verifying stats collected when comparing two directories with similar file contents but the metadata for the files have changed between snapshots but content remains unchanged. Expectation is all the relevant file level stats fields are updated.
Existing tests have been updated due to stats now being printed in addition to previous output.
Avoid starting a maintenance task on the server when
the repository connection (configuration) is read-only.
Also:
* Check for read-only repo before running maintenance.
* Move direct repo check to startMaintenanceManager.
* Rename function as maybeStartMaintenanceManager.
The name reflects that the function may not start a
maintenance task manager.
* Add connect options to repotesting.
* Add test for maintenance on read-only repo.
- Fixes: #4373
The kopia server was not uploading any logs to the repository,
because "repodiag" blob uploads would always fail.
The cause was the following: when the (log) repodiag blob
PUT operation was initiated, the `Context` used for this
operation was already canceled.
The context used for blob uploads is passed to
`repodiag.NewLogManager` when opening the repository.
In the case of the kopia server, the repository is asynchronously
opened in `server.Server.InitReposotoryAsync`. The context
passed to `repo.Open` is canceled after the "open repository"
server task completes.
This issue was introduced in #1691
Change:
Use `context.WithoutCancel()` instead of the context passed
when the repo diagnoser is created.
New tests are included to reproduce this failure and verify
the fix.
- test: ensure server logs are uploaded to the repo
- test: honor cancellation in map storage
- test: repodiag context cancellation
Ref:
- #1691
* feat(general): various notifications improvements
* added API to test notification profiles
* added --http-header to webhook notification configuration
* refactored configuration to always apply defaults before persisting options in the repository
* added 'notification profile show --profile-name=X' command
* more tests
* more test coverage
* report notification code coverage
- Fixes: #4257
- Fix: Handle case when the hash version is unset in the user profile
- Add tests for default password hash
- Fix failing authenticator test: When the password hashing version
is not set, then the default one is used, so verifying the password
works.
Refactor authenticator tests:
- Nit: rename test variable
- Check error when calling SetPassword
- Add comments for test cases
- Nit: user assert.Equal in verifyRepoAuthenticator
- Split repository authenticator tests to test invariants separately
This was caused by the client using key derivation algorithm
from a config file (which did not have it when it was generated
using old version of Kopia).
Fixes#4254
