[MXS-3063] error : Sync marker mismatch. Created: 2020-07-03  Updated: 2021-08-23  Resolved: 2021-08-20

Status: Closed
Project: MariaDB MaxScale
Component/s: avrorouter
Affects Version/s: 2.4.10
Fix Version/s: 2.5.16

Type: Bug Priority: Major
Reporter: wilsonlau Assignee: markus makela
Resolution: Fixed Votes: 0
Labels: None

Attachments: PNG File QQ截图20200703120740.png    
Sprint: MXS-SPRINT-110, MXS-SPRINT-138

 Description   

2020-07-03 10:57:41 error : Sync marker mismatch.
OOM: _func_
OOM: _func_
2020-07-03 10:57:41 error : Sync marker mismatch.
2020-07-03 10:57:41 error : Failed to read field value 'event_type', type 'enum' at file offset 135257, record number 664.
OOM: _func_
2020-07-03 10:57:41 error : Sync marker mismatch.
OOM: _func_
2020-07-03 10:57:41 error : Sync marker mismatch.
OOM: _func_
2020-07-03 10:57:41 error : Sync marker mismatch.
2020-07-03 10:57:41 error : Failed to read field value 'event_type', type 'enum' at file offset 139333, record number 664.
OOM: _func_
OOM: _func_
2020-07-03 10:57:41 error : Sync marker mismatch.
2020-07-03 10:57:41 error : Failed to read field value 'server_id', type 'int' at file offset 139356, record number 664.

 Comments   
Comment by markus makela [ 2020-07-03 ]

Looks like MaxScale ran out of memory, how high was the memory usage at the time this happens?

Comment by markus makela [ 2021-08-20 ]

The memory problem is more likely a result of requesting a value that's too large for malloc to allocate and it returns a null pointer. I suspect that this is caused by the same problem that causes the sync marker mismatch: either the file is corrupted or the file offset is wrong.

Comment by markus makela [ 2021-08-20 ]

I sort of managed to reproduce this by writing random data into the avro files. I think for now just aborting the connection immediately when corruption is detected is an adequate solution for this particular problem.

Generated at Thu Feb 08 04:18:41 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.