While fail-fast and best-effort make for good
categories at the extreme ends of error handling,
we keep finding outselves wanting to operate
in a middle ground. This change introduces a new
error handling category: FailAfterRecovery. This
option tells corso to complete as much of its
process as it can, even if it recovers from errors.
But at the end of processing, if it recovered from
any errors, an error is returned for the operation.
This behavior is the new failure handling default,
instead of failFast.
---
#### Does this PR need a docs update or release note?
- [x] ⛔ No
#### Type of change
- [x] 🌻 Feature
#### Test Plan
- [x] 💪 Manual
- [x] ⚡ Unit test
Replaces the operations graphConnector reference
with an interface. Restore and Backups have
separate, unique interfaces.
<!-- PR description-->
---
#### Does this PR need a docs update or release note?
- [x] ⛔ No
#### Type of change
- [x] 🌻 Feature
#### Issue(s)
* #2825
#### Test Plan
- [x] ⚡ Unit test
- [x] 💚 E2E
Moves the generation of graphConnector to the
NewBackupOp and NewRestoreOp constructors
inside the repository. This removes gc creation
from the backup and restore operations, and
requires a reference to gc to exist in operations.
This sets up two changes: 1/ mocking of GC
within operation (the next PR will replace it with
an interface). 2/ the ability to perform a gc
validation step within the operation construction,
and thus the capacity to look up a resource owner
display name or id, and vice versa.
---
#### Does this PR need a docs update or release note?
- [x] ⛔ No
#### Type of change
- [x] 🌻 Feature
#### Issue(s)
* #2825
#### Test Plan
- [x] ⚡ Unit test
- [x] 💚 E2E
Mostly find/replace on errors.N and errors.W. Also turns all wrapf into wrap, and removes as many errorf calls as possible.
Might follow up with a linter to enforce this change.
---
#### Does this PR need a docs update or release note?
- [x] ⛔ No
#### Type of change
- [x] 🧹 Tech Debt/Cleanup
#### Issue(s)
* #1970
#### Test Plan
- [x] ⚡ Unit test
- [x] 💚 E2E
## Description
Renaming the funcs in the fault
package to be more clear about
their purpose and behavior. Largely
just find&replace changes, except
for fault.go and the fault examples.
## Does this PR need a docs update or release note?
- [x] ⛔ No
## Type of change
- [x] 🧹 Tech Debt/Cleanup
## Issue(s)
* #1970
## Test Plan
- [x] ⚡ Unit test
- [x] 💚 E2E
## Does this PR need a docs update or release note?
- [x] ⛔ No
## Type of change
- [x] 🧹 Tech Debt/Cleanup
## Issue(s)
* #1970
## Test Plan
- [x] ⚡ Unit test
- [x] 💚 E2E
## Description
Adds the fault.Errors struct (now exported) to the
operations base. stats.Errs is retained in the
backup and restore wrappers to avoid breaking
changes and allow for deserialization. We will
continue to use the current error return until
dependencies are fully updated to use Errors.
## Does this PR need a docs update or release note?
- [x] ⛔ No
## Type of change
- [x] 🧹 Tech Debt/Cleanup
## Issue(s)
* #1970
## Test Plan
- [x] ⚡ Unit test
- [x] 💚 E2E
## Description
A couple minor refactors. First, combines the
graph/service_helper.go and graph/service.go
files into a single file (just lift'n'shift, no logic
changes). Second, renames CreateHTTPClient
to just HTTPClient. Third, replaces the new
LargeItemClient constructor with extensible
options for the http client, starting with a
NoTimeout option. This will provide longer
term maintenance over multiple constructors.
This change sets up QoL for building multiple services
and clients within api layer handlers, which will appear
in an immediately following PR.
## Does this PR need a docs update or release note?
- [x] ⛔ No
## Type of change
- [x] 🧹 Tech Debt/Cleanup
## Issue(s)
* #2299
## Test Plan
- [x] ⚡ Unit test
- [x] 💚 E2E
## Description
This is a quick hack to satisfy a primary case of PII scrubbing. We expect to revisit it in the future.
## Does this PR need a docs update or release note?
- [x] ⛔ No
## Type of change
- [x] 🌻 Feature
## Issue(s)
* #2284
## Test Plan
- [x] 💪 Manual
- [x] ⚡ Unit test
## Description
This PR reintroduces the changes from #2266 with a change to *not* reset the transport
when initializing the shared client.
Doing so was removing the retry and other middlewares
and also resulting in throttled requests being masked as success
Also - we now decorate our download traffic with an ISV tag as recommended [here](https://learn.microsoft.com/en-us/sharepoint/dev/general-development/how-to-avoid-getting-throttled-or-blocked-in-sharepoint-online#how-to-decorate-your-http-traffic)
## Does this PR need a docs update or release note?
- [ ] ✅ Yes, it's included
- [x] 🕐 Yes, but in a later PR
- [ ] ⛔ No
## Type of change
<!--- Please check the type of change your PR introduces: --->
- [ ] 🌻 Feature
- [x] 🐛 Bugfix
- [ ] 🗺️ Documentation
- [ ] 🤖 Test
- [ ] 💻 CI/Deployment
- [ ] 🧹 Tech Debt/Cleanup
## Issue(s)
<!-- Can reference multiple issues. Use one of the following "magic words" - "closes, fixes" to auto-close the Github issue. -->
* #2266
## Test Plan
<!-- How will this be tested prior to merging.-->
- [ ] 💪 Manual
- [ ] ⚡ Unit test
- [x] 💚 E2E
## Description
onedrive currently constructs a new http client
for every file it downloads. This causes the OS
to generate extra sockets, and hang onto them
after the download is complete. Replacing these
one-off clients with a singular, re-used client-
which is the behavior and standard suggested
for golang http clients- minimizes system
resource consumption.
## Does this PR need a docs update or release note?
- [x] ⛔ No
## Type of change
- [x] 🐛 Bugfix
## Issue(s)
* closes#2262
## Test Plan
- [x] 💪 Manual
- [x] ⚡ Unit test
- [x] 💚 E2E
## Description
add logging to the observe package, assume that
every instance where a message is observed, it
also gets logged.
Merger may want to wait until logging to a file is the standard behavior, else the terminal might get messy/buggy.
## Does this PR need a docs update or release note?
- [x] ⛔ No
## Type of change
- [x] 🧹 Tech Debt/Cleanup
## Issue(s)
* closes#2061
## Test Plan
- [x] 💪 Manual
## Description
Starting off changes with a minor refactor. Hopefully will
keep code cleaner as we populate the backup operation in
the coming changes.
## Type of change
- [x] 🐹 Trivial/Minor
## Issue(s)
* #1725
## Description
Adds a NoData status to operations in the event
that a backup (or, possibly in the future, restore)
finishes processing without having any items to
store or restore.
## Type of change
- [x] 🌻 Feature
## Issue(s)
* #1000
## Test Plan
- [x] 💪 Manual
- [x] ⚡ Unit test
## Description
1/ introduce an ID() method int the account
package which provides the tenantID of the
configured account provider.
2/ introduce a event bus mock for testing.
## Type of change
- [x] 🌻 Feature
## Issue(s)
* #741
## Test Plan
- [x] ⚡ Unit test
## Description
Adds events for repo init, backup start and end,
and restore start and end. Not all expected
values are available. TODOs have been marked
for values that need future implementation.
## Type of change
- [x] 🌻 Feature
## Issue(s)
* #741
## Test Plan
- [x] ⚡ Unit test
swaps the corso go module from github.com/
alcionai/corso to github.com/alcionai/corso/src
to align with the location of the go.mod and
go.sum files inside the repo.
All other changes in the repository update the
package imports to the new module path.
* Fix wsl lint errors in pkg package
* Fix wsl lint errors in most of internal package
Leave some sub-packages out that have higher churn at the moment.
In order for sdk users to build an operation using a repo,
without importing /inernal packages, the operation
options needs to be available via a /pkg package.
* store backup operation results in the backup manifest
Adds backup operation metadata like the outcome
statistics and selector definitions to the backup manifest
entry. These additional details will appear when users
call `corso backup list`.
* adds the onError option to operations
Adds the OnError option to operations.Options. OnError tells
corso whether to continue despite concurrent processing
errors, or to exit processing on any error. Also includes flag
support for setting the option. Only adds the options, does
not assert error handling behavior in corso.
* adds store package for wrapping model_store
Introduces the pkg/store package, which contains funcs
for wrapping the model_store with common requests.
This package choice was made for its combination
of being in an accessible place, centralizing functionality
and not introducing circular dependencies.
Update the backup operation to create RestorePoint and RestorePointDetails models in the repository
Add modelstore to the operation to allow backup/restore operations to update/query for corso models
Closes#268
* Split KopiaWrapper into repo handle and logic
With ModelStore, multiple structs need a reference to the kopia repo.
Make a small wrapper class (conn) that can open and initialize a repo. The
wrapper handles concurrent closes and opens and does ref counting to
ensure it only drops the kopia handle when the last reference is closed.
Rename KopiaWrapper to Wrapper and keep backup/restore functionality
in it.
e2e wiring of persistence is not yet complete.
Will need modelstore integration, and additional
information about file and error counts from kw and gc.
* add operation results structs
Operations, both backup and restore, need to hold the
results of their operation, and be able to marshal the struct
to json for output.
* hook up restore end-to-end
Now that GC and KW both provide restore operations for a
single message, we can hook up the end-to-end restore
process. Integration tests for this change will follow in the
next PR.
* introduce operations (#114)
Adds the /internal/operations package. This is a starting point
which only includes the base operation struct and handling.
Per-process (backup/restore) operations to come at a later time.
* adding comments for clarity in op_progress
