Random crashes with encrypted realm.

[TheHmmka]

How frequently does the bug occur?

All the time

Description

Application constantly crashes in random transactions.

Stacktrace & log output

uncaught exception in notifier thread: N5realm5_impl17TransactLogParser14BadTransactLogE: Bad transaction log
2022-01-24 18:16:38.391094+0300 Walken[15692:3232033] uncaught exception in notifier thread: N5realm5_impl17TransactLogParser14BadTransactLogE: Bad transaction log
libc++abi: terminating with uncaught exception of type realm::_impl::TransactLogParser::BadTransactLog: Bad transaction log
dyld4 config: DYLD_LIBRARY_PATH=/usr/lib/system/introspection DYLD_INSERT_LIBRARIES=/Developer/usr/lib/libBacktraceRecording.dylib:/Developer/usr/lib/libMainThreadChecker.dylib:/Developer/Library/PrivateFrameworks/DTDDISupport.framework/libViewDebuggerSupport.dylib
terminating with uncaught exception of type realm::_impl::TransactLogParser::BadTransactLog: Bad transaction log

Can you reproduce the bug?

Yes, always

Reproduction Steps

Version

10.21.1

What SDK flavour are you using?

Local Database only

Are you using encryption?

Yes, using encryption

Platform OS and version(s)

iOS 15

Build environment

Xcode version: 13.2.1
Dependency manager and version: CocoaPods 1.11.2

[TheHmmka]

I found next description in the Doc, may be it's a reason:

If multiple processes need to access a realm simultaneously, use an unencrypted realm. Encrypted realms cannot be accessed by multiple processes at the same time.
As an alternative, you can store data that you want to encrypt as NSData properties on realm objects. Then, you can encrypt and decrypt individual fields.

[dianaafanador3]

Hi @TheHmmka I see you can reproduce the issue, maybe you can share a repo or a code snippet so we can reproduce it in our side.
The warning on our documentation refers to trying to access a shared realm from different processes, for example a realm stored on a shared group which is used by different apps at the same time?, if this the case that may be the cause of the crash.

[TheHmmka]

Hi @TheHmmka I see you can reproduce the issue, maybe you can share a repo or a code snippet so we can reproduce it in our side.
The warning on our documentation refers to trying to access a shared realm from different processes, for example a realm stored on a shared group which is used by different apps at the same time?, if this the case that may be the cause of the crash.

Nope, this is not my case.

Sorry, I can't share production code. I'll try to create demo project and catch this issue.

[tfe]

I've started getting these too, possibly concurrently with starting to use Realm Notifications, or starting to use async/await, all of which happened around the same time.

I am not accessing the Realm from different processes in my case.

[dianaafanador3]

HI @tfe would you mind sharing your stack trace and/or the code where you het the crash, with a little more info we may be able to reproduce the issue.

[tfe]

@dianaafanador3 I haven't seen it do it in development... all I have is the exception report via Sentry, and it doesn't give very much (unless I'm not looking in the right place?):

libsystem_kernel    0x1b8efab2c  __workq_kernreturn
libsystem_pthread   0x1f25fe13c  _pthread_wqthread

However from their breadcrumbs I can see that it's likely happening during the same operation I mentioned and shared code for over in #7641. It makes sense that that operation would surface issues since it's the most extensive exercise the database receives in the lifecycle of this app.

[sainttail]

@TheHmmka Got this error when upgrading from realm 10.20.0 to 10.21.1 on iOS 15 both simulator and real device.

Have you try going back to use 10.20.2?
It does not crash for me.

[TheHmmka]

@TheHmmka Got this error when upgrading from realm 10.20.0 to 10.21.1 on iOS 15 both simulator and real device.

Have you try going back to use 10.20.2? It does not crash for me.

I'll try, thank you!

[Saul-Liang]

We're also seeing this crash on 10.22.0. Downgrading to 10.20.2 seems to fix the issue for us - thanks @sainttail 🙏

[bfichter]

We're also experiencing this sporadically after upgrading to 10.21.1 with encryption enabled.

[leemaguire]

@bfichter @Saul-Liang can you attach any stack traces?

[dianaafanador3]

This issue may be related to this realm/realm-core#5174

[driq]

What we know so far:

• The issue appears in Realm-swift v10.21.1, per Random crashes with encrypted realm. #7629 (comment)
• Realm-swift 10.20.2 works fine, as reported in Random crashes with encrypted realm. #7629 (comment)
• The same issue is also reported in Decryption crash at realm version 10.12.0 realm-js#4428, suggesting that the issue is in realm-core (registered as exposure of partial writes with encryption (and multiprocess) realm-core#5174)
• In Realm-swift v10.21.1, Realm-core was upgraded from 11.6.1 to `11.8.0.

My conclusion:

• The issue was introduced in Realm core v11.7.0 or v11.8.0
• Realm-swift v10.21.0 still contained Realm-core 11.6.1, and should work without issues if the above is true

What to try:

• Downgrading Realm-swift to v10.21.0 should fix the issue until exposure of partial writes with encryption (and multiprocess) realm-core#5174 is fixed

PS. I did not expect such relatively major changes in a patch release of Realm-swift, especially considering that Realm core 11.8.0 contains breaking changes (according to it's own changelog). Is there any reason not to adhere to semver?

[sync-by-unito]

➤ Finn Andersen commented:

Fixed by realm-core 11.14

[TheHmmka]

Now encryption works as expected.