Re: replication problem / myisamchk: error

[olli]

mysql has a medium load: (4.1.12log) up 6+00:00:07 [15:09:52]
Queries: 109.4M qps: 221 Slow: 1.1k Se/In/Up/De(%): 16/06/01/00
qps now: 416 Slow qps: 0.0 Threads: 690 ( 6/ 178) 01/05/00/00
Cache Hits: 263.1k Hits/s: 0.5 Hits now: 0.0 Ratio: 1.5% Ratio now: 0.0%
Key Efficiency: 99.9% Bps in/out: 4.1k/ 5.5k Now in/out: 17.8k/52.5k

- all slaves are running with "read-only"

the mess up shows if we reload a slave with "load data from master" - than the error log of this slave show a lot of different errors for different tables - if we check the same tables of the master after that (with mysql shutdown before) we see that they are "should be repaired"..

this are the errors in the slave logs - the appear just one time after reloading data from master:

-
050617 15:11:56 [ERROR] Slave I/O thread exiting, read up to log 'db-master-bin.000170', position 92194474
050617 15:11:56 [ERROR] Error reading relay log event: slave SQL thread was killed
050617 15:12:00 [ERROR] Delete link points outside datafile at 1084
050617 15:12:00 [Note] Retrying repair of: './cmsl/anbieter' with keycache
050617 15:12:00 [ERROR] Delete link points outside datafile at 1084
050617 15:12:00 [ERROR] Delete link points outside datafile at 3696
050617 15:12:00 [ERROR] Delete link points outside datafile at 6136
050617 15:12:00 [ERROR] Delete link points outside datafile at 9200
050617 15:12:00 [ERROR] Delete link points outside datafile at 16692
050617 15:12:00 [ERROR] Delete link points outside datafile at 19748
050617 15:12:00 [ERROR] Couldn't fix table with quick recovery: Found wrong number of deleted records
050617 15:12:00 [ERROR] Run recovery again without -q
050617 15:12:00 [Note] Retrying repair of: './cmsl/beitrag' with keycache

-
050617 15:12:09 [Note] Retrying repair of: './cms/tarif_rate_costpoint' with keycache
050617 15:12:09 [ERROR] Delete link points outside datafile at 45340
050617 15:12:09 [ERROR] Delete link points outside datafile at 56284
050617 15:12:09 [ERROR] Delete link points outside datafile at 67552
050617 15:12:09 [ERROR] Delete link points outside datafile at 67752
050617 15:12:09 [ERROR] Delete link points outside datafile at 85400
050617 15:12:09 [ERROR] Found block that points outside data file at 85424
050617 15:12:09 [ERROR] Found block that points outside data file at 85464
050617 15:12:09 [ERROR] Delete link points outside datafile at 86664
050617 15:12:09 [ERROR] Found block that points outside data file at 86688
050617 15:12:09 [ERROR] Found block that points outside data file at 86728
050617 15:12:09 [ERROR] Delete link points outside datafile at 90896
050617 15:12:09 [ERROR] Found block that points outside data file at 90920
050617 15:12:09 [ERROR] Found block that points outside data file at 90960
050617 15:12:09 [ERROR] Delete link points outside datafile at 92324
050617 15:12:09 [ERROR] Found block that points outside data file at 92348
050617 15:12:09 [ERROR] Found block that points outside data file at 92388
050617 15:12:09 [ERROR] Found block that points outside data file at 92428
050617 15:12:09 [ERROR] Found block that points outside data file at 92468
050617 15:12:09 [ERROR] Found block that points outside data file at 92508
050617 15:12:09 [ERROR] Delete link points outside datafile at 93588
050617 15:12:09 [ERROR] Found block that points outside data file at 93612
050617 15:12:09 [ERROR] Found block that points outside data file at 93652
050617 15:12:09 [ERROR] Delete link points outside datafile at 120520
050617 15:12:09 [ERROR] Delete link points outside datafile at 122996

-
050617 15:12:34 [Note] Retrying repair of: './cms/reminder_addresses' with keycache
050617 15:12:34 [ERROR] Delete link points outside datafile at 864
050617 15:12:34 [ERROR] Delete link points outside datafile at 1248
050617 15:12:34 [ERROR] Delete link points outside datafile at 0
050617 15:12:34 [Note] Retrying repair of: './cms/tarifrechner_stats' with keycache
050617 15:12:34 [ERROR] Delete link points outside datafile at 0
050617 15:12:34 [ERROR] Delete link points outside datafile at 32516
050617 15:12:34 [ERROR] Delete link points outside datafile at 34272
050617 15:12:34 [ERROR] Delete link points outside datafile at 34576
050617 15:12:34 [ERROR] Couldn't fix table with quick recovery: Found wrong number of deleted records
050617 15:12:34 [ERROR] Run recovery again without -q
050617 15:12:34 [Note] Retrying repair of: './cms/teaser_rotation_stats' with keycache
050617 15:12:36 [ERROR] Couldn't fix table with quick recovery: Found wrong number of deleted records
050617 15:12:36 [ERROR] Run recovery again without -q
050617 15:12:36 [Note] Retrying repair of: './cms/teaser_rotation_stats' with keycache
050617 15:12:37 [ERROR] Couldn't fix table with quick recovery: Found wrong number of deleted records
050617 15:12:37 [ERROR] Run recovery again without -q
050617 15:12:37 [Note] Retrying repair of: './cms/teaser_rotation_stats' with keycache
050617 15:19:25 [Note] Slave SQL thread initialized, starting replication in log 'db-master-bin.000171' at position 15726, relay log './db-slave3-relay-bin.0
00001' position: 4
050617 15:19:25 [Note] Slave I/O thread: connected to master '[e-mail]:3306', replication started in log 'db-master-bin.000171' at posit
ion 15726

if we make a "check table" on the master now for all this tables and then again "load data from master" on the slaves everthing is ok for some hours again - but a little time later we have them again in the logs after the next data reload..