version: 1.2.0
this is difficult to reproduce, as I guess the issue is related to data race or monotone losing track of the previous WAL and after being unable to delete the old ones.
during active writing, monotone usually manages to create and clean up old WALs.
but sometimes I see endless growth of WALs and checkpoint doesn't help.
the only way to fix this is to reopen repository. monotone will attempt to recover from WALs (I see in log wal: 99.2 million records processed) and then checkpoint will help and old WALs will be deleted.
the following show commands were executed after the checkpoint (before restart).
// show config
{
"version": "1.2.0",
"uuid": "3f10cafb-b46d-2069-dfa6-bb3663b560a5",
"directory": "./data",
"log": true,
"log_to_file": true,
"log_to_stdout": false,
"mm_page_size": 2097152,
"mm_limit": false,
"mm_limit_wm": 8589934592,
"mm_limit_behaviour": "block",
"mm_cache": 8589934592,
"wal": true,
"wal_crc": true,
"wal_sync_on_create": true,
"wal_sync_on_close": true,
"wal_sync_on_write": false,
"wal_sync_interval": 1000,
"wal_size": 104857600,
"wal_recover": false,
"serial": true,
"interval": 3000000,
"compression": "zstd",
"compression_level": 0,
"workers": 3,
"workers_upload": 1,
"ssn": 501000301,
"lsn": 5100031
}
// show wal
{
"active": true,
"crc": true,
"sync_on_create": true,
"sync_on_close": true,
"sync_on_write": false,
"sync_interval": 1000,
"size": 104857600,
"recover": false,
"lsn": 5100031,
"lsn_min": 4107989,
"files": 16
}
// show storages
{
"main": {
"config": {
"uuid": "3f10cafb-b46d-2069-dfa6-bb3663b560a5",
"name": "main",
"path": "",
"cloud": "",
"cloud_drop_local": true,
"sync": true,
"crc": false,
"refresh_wm": 41943040,
"region_size": 131072,
"compression": "zstd",
"compression_level": 0,
"encryption": ""
},
"stats": {
"min": 0,
"max": 503999999,
"partitions": 168,
"partitions_local": 168,
"partitions_cloud": 0,
"regions": 254959,
"events": 501000301,
"events_heap": 0,
"size": 5279259823,
"size_uncompressed": 33457410144,
"size_heap": 0
}
}
}
// show partitions (last ones)
...
}, {
"min": 498000000,
"max": 500999999,
"events": 3000000,
"events_heap": 0,
"regions": 458,
"size": 16405648,
"size_uncompressed": 60034938,
"size_heap": 0,
"created": 1766138030146043,
"refreshed": 1766138030880078,
"refreshes": 2,
"on_storage": true,
"on_cloud": false,
"storage": "main"
}, {
"min": 501000000,
"max": 503999999,
"events": 301,
"events_heap": 0,
"regions": 1,
"size": 1822,
"size_uncompressed": 6226,
"size_heap": 0,
"created": 1766138030879567,
"refreshed": 1766138334499597,
"refreshes": 1,
"on_storage": true,
"on_cloud": false,
"storage": "main"
}]
version: 1.2.0
this is difficult to reproduce, as I guess the issue is related to data race or monotone losing track of the previous WAL and after being unable to delete the old ones.
during active writing, monotone usually manages to create and clean up old WALs.
but sometimes I see endless growth of WALs and checkpoint doesn't help.
the only way to fix this is to reopen repository. monotone will attempt to recover from WALs (I see in log wal: 99.2 million records processed) and then checkpoint will help and old WALs will be deleted.
the following show commands were executed after the checkpoint (before restart).