Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
Home -> Community -> Usenet -> c.d.o.misc -> Oracle Replication and Locking
Howdy everyone,
We are currently having problem asynch replication locking when the
network
is down. Hopefully someone out there will have some suggestions on what
we can do.
We are trying to test a replicated system's recovery ability when the
network is
down, and we ran across an interesting case when the system is
disconnected in
the middle of a snapshot update back to a master database.
Here's the results of the tests we performed here on the machines.
Scenario 1
Scenario 2
Here's a listing of the locks:
Username Session Serial Lock Resource Resource Mode Mode ID Number Type ID 1 ID 2 Held Req FUEL 7 36385 DX 13 0 NULL NONE FUEL 29 18637 JQ 0 2030 X NONE FUEL 11 19922 JQ 0 2032 X NONE FUEL 8 36339 JQ 0 2035 X NONE FUEL 10 36355 JQ 0 2031 X NONE FUEL 10 36355 DX 10 0 NULL NONE FUEL 29 18637 TM 119 0 SS NONE FUEL 7 36355 TM 4765 0 X NONE FUEL 10 36355 TM 4788 0 X NONE FUEL 10 36355 TX 196647 155706 X NONE FUEL 29 18637 TX 131107 155892 X NONE FUEL 8 36339 TX 262147 155291 X NONE FUEL 11 19922 TX 262161 155290 X NONE FUEL 10 36355 TM 119 0 SS NONE FUEL 8 36339 TM 4816 0 X NONE FUEL 11 19922 TM 119 0 SS NONE FUEL 7 36385 TM 4767 0 X NONE FUEL 7 36385 TX 327746 154971 X NONE FUEL 8 36339 TM 119 0 SS NONE FUEL 29 18637 TM 4777 0 X NONE FUEL 29 18637 TM 4774 0 X NONE FUEL 7 36385 TM 119 0 RX NONE FUEL 10 35355 TM 4785 0 X NONE FUEL 8 36339 TM 4813 0 X NONE ROOT 12 34 TM 4785 0 NONE RX
Object ID's
119 SNAP$ 4765 SNAP$_ROAD_CALL 4788 USLOG$_TRANSACT 4816 USLOG$_TANK 4767 USLOG$_ROAD_CALL 4777 USLOG$_MTOTRANS 4774 SNAP$_MTOTRANS 4785 SNAP$_TRANSACT 4813 SNAP$_TANK
The problem table/snapshot is "TRANSACT", which is what the last line of the preceding output shows that it's waiting for a RX lock, but it can't because the line two lines up shows that a process has it in X mode. BTW, the "FUEL" user is where we have our replication defined, and it running. The "ROOT" user is the Oracle user where our background daemons are logged into.
I guess my questions for this behavior would be:
Thanks for any info!
--Gary
--Received on Wed Jan 08 1997 - 00:00:00 CST
---------------------------+----------------------------------------------
Gary Eckhardt | "in this day & age...music performed by Database Consultants, Inc. | humans...hum!?" --wilde silas tomkyn dcigary_at_txdirect.net | R,DW,HAHB! gary_eckhardt_at_realworld.com| R^3 = "Real World. Real Smart. Real Quick." (210)344-6566 | http://www.realworld.com/