TrueCloudLab/frostfs-node
20
1
Fork 27
Code Issues 105 Pull requests 7 Projects Releases 2 Activity Actions

[#139] test: Add test storage implementation #173

Merged
fyrchik merged 1 commit from ale64bit/frostfs-node:fix/139-unit_test_storage into master 2023-03-29 14:28:50 +00:00
Conversation 3 Commits 1 Files changed 20 +617 -208
ale64bit commented 2023-03-24 09:27:13 +00:00
Member
Copy link

This aims to reduce the usage of chmod hackery to induce or simulate
OS-related failures.

Signed-off-by: Alejandro Lopez a.lopez@yadro.com

This aims to reduce the usage of chmod hackery to induce or simulate OS-related failures. Signed-off-by: Alejandro Lopez <a.lopez@yadro.com>
ale64bit force-pushed fix/139-unit_test_storage from f6b6097f0c to 4907d78dab 2023-03-24 09:31:15 +00:00 Compare
requested reviews from storage-core-committers, storage-core-developers 2023-03-24 09:31:44 +00:00
dstepanov-yadro approved these changes 2023-03-28 06:26:45 +00:00
dstepanov-yadro commented 2023-03-28 06:27:27 +00:00
Member
Copy link

It seems that it was easier to use gomockery :-)

It seems that it was easier to use gomockery :-)
ale64bit commented 2023-03-28 10:38:07 +00:00
Author
Member
Copy link

It seems that it was easier to use gomockery :-)

I agree :))

> It seems that it was easier to use gomockery :-) I agree :))
ale64bit force-pushed fix/139-unit_test_storage from 4907d78dab to e5c5235357 2023-03-29 07:25:16 +00:00 Compare
ale64bit commented 2023-03-29 07:26:11 +00:00
Author
Member
Copy link

(force push only for fixing go.mod, go.sum conflicts)

(force push only for fixing `go.mod`, `go.sum` conflicts)
requested review from dstepanov-yadro 2023-03-29 07:26:19 +00:00
dstepanov-yadro approved these changes 2023-03-29 07:51:44 +00:00
fyrchik reviewed 2023-03-29 11:28:01 +00:00
go.mod Outdated
@ -30,3 +30,3 @@
github.com/spf13/pflag v1.0.5
github.com/spf13/viper v1.15.0
github.com/stretchr/testify v1.8.1
github.com/stretchr/testify v1.8.2
fyrchik commented 2023-03-29 11:27:57 +00:00
Owner
Copy link

Can we have a separate commit for this?

Can we have a separate commit for this?
ale64bit commented 2023-03-29 11:49:54 +00:00
Author
Member
Copy link

done

done
fyrchik commented 2023-03-29 13:55:18 +00:00
Owner
Copy link

We still have some changes in go.sum

We still have some changes in `go.sum`
ale64bit commented 2023-03-29 14:26:40 +00:00
Author
Member
Copy link

hm, I ran go mod tidy. I don't really touch these files manually. Maybe it's ok now.

hm, I ran `go mod tidy`. I don't really touch these files manually. Maybe it's ok now.
fyrchik marked this conversation as resolved
pkg/local_object_storage/blobstor/blobstor_test.go Outdated
@ -16,3 +16,1 @@
const blobovniczaDir = "blobovniczas"
func defaultStorages(p string, smallSizeLimit uint64) []SubStorage {
func defaultTestStorages(p string, smallSizeLimit uint64) ([]SubStorage, *teststore.TestStore, *teststore.TestStore) {
fyrchik commented 2023-03-29 10:14:51 +00:00
Owner
Copy link

Do we need to return the last 2 arguments? Why not casting a type from SubStorage?

Do we need to return the last 2 arguments? Why not casting a type from `SubStorage`?
ale64bit commented 2023-03-29 11:51:29 +00:00
Author
Member
Copy link

What's wrong with returning both arguments? Seems to me that way it's more clear what is returned and the cast/indexing is unnecessary (those assume a type and slice size).

What's wrong with returning both arguments? Seems to me that way it's more clear what is returned and the cast/indexing is unnecessary (those assume a type and slice size).
fyrchik commented 2023-03-29 13:50:56 +00:00
Owner
Copy link

That the number of returned arguments directly depends on the slice length, and casting is not that hard.

That the number of returned arguments directly depends on the slice length, and casting is not that hard.
fyrchik commented 2023-03-29 13:55:37 +00:00
Owner
Copy link

I don't insist, just share an opinion.

I don't insist, just share an opinion.
dstepanov-yadro commented 2023-03-29 14:07:46 +00:00
Member
Copy link

Seems to me it's more clear to return struct with named fields. Casting is evil and should be prohibited.

Seems to me it's more clear to return struct with named fields. Casting is evil and should be prohibited.
😆 1
fyrchik marked this conversation as resolved
pkg/local_object_storage/blobstor/teststore/teststore.go Outdated
@ -0,0 +55,4 @@
switch {
case s.overrides.Open != nil:
return s.overrides.Open(readOnly)
case s.st != nil:
fyrchik commented 2023-03-29 11:18:59 +00:00
Owner
Copy link

Сan we enforce this in new and make this branch a default? Implicit panic is also not bad here.

Сan we enforce this in `new` and make this branch a `default`? Implicit panic is also not bad here.
ale64bit commented 2023-03-29 11:52:52 +00:00
Author
Member
Copy link

This is intentional, so that a TreeStore can also be used purely as a mock (e.g. to verify only certain calls are executed and nothing is propagated to any underlying storage).
Re. the implicit panic, isn't better an explicit panic with specific error message rather than a nil dereference panic which might not be immediately clear?

This is intentional, so that a `TreeStore` can also be used purely as a mock (e.g. to verify only certain calls are executed and nothing is propagated to any underlying storage). Re. the implicit panic, isn't better an explicit panic with specific error message rather than a nil dereference panic which might not be immediately clear?
👍 1
fyrchik marked this conversation as resolved
pkg/local_object_storage/shard/control_test.go Outdated
@ -43,0 +53,4 @@
var allowedMode atomic.Int64
openFileMetabase := func(p string, f int, perm fs.FileMode) (*os.File, error) {
if int64(f&3) == allowedMode.Load() {
fyrchik commented 2023-03-29 11:24:24 +00:00
Owner
Copy link

3 looks like a magic here, can we define a constant as os.O_RDONLY | os.O_RDWR | os.O_WRONLY? Or somehow refer to them in the comment for the named constant.

`3` looks like a magic here, can we define a constant as `os.O_RDONLY | os.O_RDWR | os.O_WRONLY`? Or somehow refer to them in the comment for the named constant.
ale64bit commented 2023-03-29 11:50:04 +00:00
Author
Member
Copy link

done

done
fyrchik marked this conversation as resolved
pkg/local_object_storage/shard/control_test.go Outdated
@ -43,0 +56,4 @@
if int64(f&3) == allowedMode.Load() {
return os.OpenFile(p, f, perm)
}
return nil, teststore.ErrDiskExploded
fyrchik commented 2023-03-29 11:25:57 +00:00
Owner
Copy link

Maybe syscall.EPERM or fs.ErrPermission?

Maybe `syscall.EPERM` or `fs.ErrPermission`?
ale64bit commented 2023-03-29 11:53:59 +00:00
Author
Member
Copy link

done. Is there any value in returning an error that is guaranteed to not be specially handled or better to use a relevant error for this path?

done. Is there any value in returning an error that is guaranteed to _not_ be specially handled or better to use a relevant error for this path?
fyrchik marked this conversation as resolved
pkg/local_object_storage/writecache/flush_test.go Outdated
@ -209,3 +210,3 @@
_, err = os.Stat(p) // sanity check
require.NoError(t, err)
require.NoError(t, os.Chmod(p, 0))
require.NoError(t, os.Truncate(p, 0))
fyrchik commented 2023-03-29 11:27:06 +00:00
Owner
Copy link

What do we achieve with Truncate here?

What do we achieve with `Truncate` here?
ale64bit commented 2023-03-29 11:55:14 +00:00
Author
Member
Copy link

os.Chmod was used here to cause the file to not be read properly, but the same can be achieved with os.Truncate so that it can't be unmarshalled. I thought at least os.Truncate didn't depend on the user running the test.

`os.Chmod` was used here to cause the file to not be read properly, but the same can be achieved with `os.Truncate` so that it can't be unmarshalled. I thought at least `os.Truncate` didn't depend on the user running the test.
fyrchik commented 2023-03-29 13:54:07 +00:00
Owner
Copy link

I mean the purpose is not obvious to me, as a code reader. May be just write "corrupted data" in the file?

I mean the purpose is not obvious to me, as a code reader. May be just write `"corrupted data"` in the file?
ale64bit commented 2023-03-29 14:26:13 +00:00
Author
Member
Copy link

ah I see. Yes, good point. I added a comment about it on this line.

ah I see. Yes, good point. I added a comment about it on this line.
fyrchik marked this conversation as resolved
ale64bit force-pushed fix/139-unit_test_storage from e5c5235357 to dd99c28fae 2023-03-29 11:49:32 +00:00 Compare
ale64bit force-pushed fix/139-unit_test_storage from dd99c28fae to 6df9f44cfd 2023-03-29 14:25:48 +00:00 Compare
requested review from dstepanov-yadro 2023-03-29 14:27:07 +00:00
fyrchik approved these changes 2023-03-29 14:28:38 +00:00
fyrchik merged commit 341fe1688f into master 2023-03-29 14:28:50 +00:00
Sign in to join this conversation.
Reviewers
No reviewers
dstepanov-yadro
fyrchik
Labels
Clear labels   
P0

Highest

  
P1

High

  
P2

Medium

  
P3

Low

  
badger

Badger DB related issues

  
frostfs-adm

   
frostfs-cli

   
frostfs-ir

   
frostfs-lens

   
frostfs-node

   
good first issue

Good for newcomers

  
triage
  
Infrastructure

Infrastructure, Automation, CI/CD or DevOps related stuff

  
blocked

The issue or PR is blocked by something

  
bug

Something is not working

  
config

Configuration format update or breaking change

  
discussion

Talking about something in order to reach a decision or to exchange ideas

  
documentation

Documentation related

  
duplicate

This issue or pull request already exists

  
enhancement

New feature

  
go

Language features adoption

  
help wanted

Extra attention is needed

  
internal

   
invalid

Something is wrong

  
kludge

This is a kludge and we know it

  
observability

Related to metrics, tracing and logs

  
perfomance

Perfomance related

  
question

More information is needed

  
refactoring

Refactoring

  
wontfix

This won't be fixed

No labels
P0
P1
P2
P3
badger
frostfs-adm
frostfs-cli
frostfs-ir
frostfs-lens
frostfs-node
good first issue
triage
Infrastructure
blocked
bug
config
discussion
documentation
duplicate
enhancement
go
help wanted
internal
invalid
kludge
observability
perfomance
question
refactoring
wontfix
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
3 participants
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference: TrueCloudLab/frostfs-node#173
No description provided.
Delete branch "ale64bit/frostfs-node:fix/139-unit_test_storage"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?