04:24:20.836 localhost-0: 04:24:20.836 op: read lun: /tmp/jfs/vdb.1_2.dir/vdb_f0003.file lba: 14290944 0x00DA1000 xfer: 4096 errno: EIO: 'I/O error'
04:24:21.338
04:24:21.338 'data_errors=1' requested. Abort rd=rd1 after last error.
04:24:21.338
java.lang.RuntimeException: 'data_errors=1' requested. Abort rd=rd1 after last error.
at Vdb.common.failure(common.java:350)
at Vdb.common.failure(common.java:297)
at Vdb.ErrorLog.countErrorsOnMaster(ErrorLog.java:224)
at Vdb.SlaveOnMaster.processSlave(SlaveOnMaster.java:207)
at Vdb.SlaveOnMaster.run(SlaveOnMaster.java:42)
It's not picked to v1.1.3 since it only happens under heavy random writes, and can be fixed by juicefs gc --compact or the new juicefs compact command.
https://github.com/juicedata/juicefs/actions/runs/9608104779