Crash in brave!AccountTrackerService::Initialize (Linux only)

[bsclifton]

Description

Going through crash reports, I'm seeing a LOT of these on Linux. Thousands.

Unclear what the impact is to users

A few examples here:
https://stats.brave.com/dashboard#crash/5d8cbd539557ce001f93a373
https://stats.brave.com/dashboard#crash/5d8cbcf6ed6e87001fe15623
https://stats.brave.com/dashboard#crash/5d8cbce5f686a0001f52aa67
https://stats.brave.com/dashboard#crash/5d8cbbf79557ce001f939cd5
https://stats.brave.com/dashboard#crash/5d8cbb3911ca47001f18d670

Call stack looks like this:

Thread 0 (crashed)
 0  brave!AccountTrackerService::Initialize(PrefService*, base::FilePath) [vector : 656 + 0x0]
    rax = 0x00007ffdd70719a8   rdx = 0x000000000000006b
    rcx = 0x000000000000000c   rbx = 0x0000000000000000
    rsi = 0x00002d1c3caaee00   rdi = 0x00002d1c3ca4b908
    rbp = 0x00007ffdd7071b30   rsp = 0x00007ffdd7071960
     r8 = 0x0000000000000000    r9 = 0x00002d1c3c9e7b18
    r10 = 0x0000000000fc0760   r11 = 0x000000000000c5c9
    r12 = 0x00007ffdd7071a20   r13 = 0x00007ffdd7071a20
    r14 = 0x00002d1c3c414b40   r15 = 0x00007ffdd7071a20
    rip = 0x00005566973141fd
    Found by: given as instruction pointer in context
 1  brave!signin::BuildIdentityManager(signin::IdentityManagerBuildParams*) [identity_manager_builder.cc : 53 + 0xe]
    rbp = 0x00007ffdd7071c60   rsp = 0x00007ffdd7071b40
    rip = 0x000055669731001e
    Found by: previous frame's frame pointer
 2  brave!IdentityManagerFactory::BuildServiceInstanceFor(content::BrowserContext*) const [identity_manager_factory.cc : 135 + 0x5]
    rbp = 0x00007ffdd7071d40   rsp = 0x00007ffdd7071c70
    rip = 0x0000556698f7af69
    Found by: previous frame's frame pointer
 3  brave!BrowserContextKeyedServiceFactory::BuildServiceInstanceFor(void*) const [browser_context_keyed_service_factory.cc : 98 + 0x9]
    rbp = 0x00007ffdd7071d60   rsp = 0x00007ffdd7071d50
    rip = 0x000055669afcc73c
    Found by: previous frame's frame pointer
 4  brave!KeyedServiceFactory::GetServiceForContext(void*, bool) [keyed_service_factory.cc : 77 + 0x9]
    rbp = 0x00007ffdd7071df0   rsp = 0x00007ffdd7071d70
    rip = 0x000055669a085bd6
    Found by: previous frame's frame pointer
 5  brave!GAIAInfoUpdateService::GAIAInfoUpdateService(Profile*) [gaia_info_update_service.cc : 43 + 0x5]
    rbp = 0x00007ffdd7071e40   rsp = 0x00007ffdd7071e00
    rip = 0x0000556698f2231e
    Found by: previous frame's frame pointer
 6  brave!GAIAInfoUpdateServiceFactory::BuildServiceInstanceFor(content::BrowserContext*) const [gaia_info_update_service_factory.cc : 40 + 0xb]
    rbp = 0x00007ffdd7071e60   rsp = 0x00007ffdd7071e50
    rip = 0x0000556698f2207e
    Found by: previous frame's frame pointer
 7  brave!BrowserContextKeyedServiceFactory::BuildServiceInstanceFor(void*) const [browser_context_keyed_service_factory.cc : 98 + 0x9]
    rbp = 0x00007ffdd7071e80   rsp = 0x00007ffdd7071e70
    rip = 0x000055669afcc73c
    Found by: previous frame's frame pointer
 8  brave!KeyedServiceFactory::GetServiceForContext(void*, bool) [keyed_service_factory.cc : 77 + 0x9]
    rbp = 0x00007ffdd7071f10   rsp = 0x00007ffdd7071e90
    rip = 0x000055669a085bd6
    Found by: previous frame's frame pointer
<SNIP>

Steps to Reproduce

?

Brave version (brave://version info)

Not a new issue; I saw some clients still on Chromium 73 versions of Brave running into this.

[fmarier]

I wonder whether tha'ts the crash behind #6054

[bsclifton]

@fmarier this has been a problem for a super long time - seems to have been present since at least Chromium 73 (from crash reports)

I'm not quite sure what the impact is to users - is this a renderer crash? From the call stack, it seems to be a hard browser process crash... but I've never ran into it 🤷‍♂

[fmarier]

I ran into it twice this morning:

Uploaded Crash Report ID 15c183a8833b7b98 (Local Context: Brave)
Crash report uploaded on Wednesday, 2 October 2019 at 10:45:35

Uploaded Crash Report ID e63765abf482eeb3 (Local Context: Brave)
Crash report uploaded on Wednesday, 2 October 2019 at 10:37:02

but I didn't really notice anything.

[fmarier]

And some more:

Uploaded Crash Report ID ae5faabfc556e1c8 (Local Context: Brave)
Crash report uploaded on Wednesday, 2 October 2019 at 13:08:18

Uploaded Crash Report ID 882bd2ad0bb56e69 (Local Context: Brave)
Crash report uploaded on Wednesday, 2 October 2019 at 12:34:56

Uploaded Crash Report ID a53060390db5aadf (Local Context: Brave)
Crash report uploaded on Wednesday, 2 October 2019 at 12:04:13

Uploaded Crash Report ID 9534a08f3ff89ed9 (Local Context: Brave)
Crash report uploaded on Wednesday, 2 October 2019 at 11:57:31

Uploaded Crash Report ID b61ad65bd0ef0407 (Local Context: Brave)
Crash report uploaded on Wednesday, 2 October 2019 at 11:55:18

Uploaded Crash Report ID 156945c0784363b1 (Local Context: Brave)
Crash report uploaded on Wednesday, 2 October 2019 at 10:47:36

It seems to happen constantly but without being noticeable by the user. I'll restart my browser in a terminal to see if this correlates with console output too.

[fmarier]

This is extremely frequent, no wonder we're getting flooded with these reports:

Uploaded Crash Report ID cf857b94133449bb (Local Context: Brave)
Crash report uploaded on Wednesday, 2 October 2019 at 14:14:50

Uploaded Crash Report ID 6c312cf14d77fb6d (Local Context: Brave)
Crash report uploaded on Wednesday, 2 October 2019 at 14:12:58

Uploaded Crash Report ID 2291def187f5d272 (Local Context: Brave)
Crash report uploaded on Wednesday, 2 October 2019 at 13:56:59

Reproduction steps on my laptop are simply: leave the browser open for a few minutes.

There doesn't appear to be anything in the terminal other than the flood of sqlite errors from #6269 .

[iefremov]

Same here. This is quite strange since the browser actually continues to work... Can't even guess which process is crashing.

[iefremov]

No traces of this since C80