[erlang-questions] Mnesia inconsistency with R11B-1

Dan Gudmundsson <>
Fri Sep 29 14:43:29 CEST 2006


Ouch, how did you manage to do that...

Call mnesia_lib:coredump() on both appservers. zip and email me privatly.

Probably it's to late to find something but I will have a look at it.

/Dan

Fredrik Thulin writes:
 > Hi
 > 
 > I upgraded from R10B-10 to R11B-1 about a week ago. Today I got a 
 > problem report that seems to be because I have inconsistencies in my 
 > Mnesia table 'phone'.
 > 
 > On node #1, a record is present that is not present on node #2 :
 > 
 >   ()49> rpc:call('', phone,  
 >      get_sipuser_locations, ["username"]).
 >   {ok,[{siplocationdb_e,{sipurl,"sip",
 >      ...
 >                         1159532776}]}
 > 
 >   ()50> rpc:call('', phone, 
 >      get_sipuser_locations, ["username"]).
 >   {ok,[]}
 >   ()51>  
 > 
 > This table is written to from at least four different nodes, but always 
 > using transactions and never with dirty writes. What should I do to 
 > figure out the cause of this? I will have to restart the nodes at the 
 > latest before Monday morning to get everything working again, but if 
 > you guys are quick I can do diagnostics on the running system =).
 > 
 > It seems to me that there is no partitioned network (both nodes list the 
 > same nodes as 'running').
 > 
 > Another indicator of the differences, and also mnesia:info() output from 
 > the two nodes :
 > 
 > ()55> length( rpc:call('', db_util, 
 > tab_to_list, [phone]) ).
 > 63
 > ()56> length( rpc:call('', db_util, 
 > tab_to_list, [phone]) ).
 > 58
 > ()57> rpc:call('', mnesia, info, []).
 > ---> Processes holding locks <---
 > ---> Processes waiting for locks <---
 > ---> Participant transactions <---
 > ---> Coordinator transactions <---
 > ---> Uncertain transactions <---
 > ---> Active tables <---
 > schema         : with 10       records occupying 1569     words of mem
 > call           : with 0        records occupying 279      words of mem
 > ===> System info in version "4.3.2", debug level = none <===
 > opt_disc. Directory "/var/yxa/mnesia/appserver" is NOT used.
 > use fallback at restart = false
 > running db nodes   = 
 > ['','','','','','','','','','','']
 > stopped db nodes   = ['','']
 > master node tables = []
 > remote             = [cpl_script_graph,
 >                       eventdata,
 >                       forward,
 >                       gruu,
 >                       numbers,
 >                       phone,
 >                       regexproute,
 >                       user]
 > ram_copies         = [call,schema]
 > disc_copies        = []
 > disc_only_copies   = []
 > [{'',ram_copies},
 > {'',ram_copies}] = [call]
 > [{'',ram_copies},
 >  {'',ram_copies},
 >  {'',ram_copies},
 >  {'',ram_copies},
 >  {'',disc_copies},
 >  {'',disc_copies},
 >  {'',ram_copies},
 >  {'',ram_copies},
 >  {'',ram_copies},
 >  {'',ram_copies},
 >  {'',ram_copies}] = [schema]
 > [{'',ram_copies},
 >  {'',ram_copies},
 >  {'',disc_copies},
 >  {'',disc_copies}] = [eventdata]
 > [{'',disc_copies},
 >  {'',disc_copies}] = [gruu,
 >                                               cpl_script_graph,
 >                                               regexproute,
 >                                               phone,
 >                                               forward,
 >                                               numbers,
 >                                               user]
 > 93 transactions committed, 0 aborted, 9 restarted, 0 logged to disc
 > 0 held locks, 0 in queue; 0 local transactions, 0 remote
 > 0 transactions waits for other nodes: []
 > ok
 > ()58> rpc:call('', mnesia, info, []).
 > ---> Processes holding locks <---
 > ---> Processes waiting for locks <---
 > ---> Participant transactions <---
 > ---> Coordinator transactions <---
 > ---> Uncertain transactions <---
 > ---> Active tables <---
 > schema         : with 10       records occupying 1569     words of mem
 > call           : with 0        records occupying 279      words of mem
 > ===> System info in version "4.3.2", debug level = none <===
 > opt_disc. Directory "/var/yxa/mnesia/appserver" is NOT used.
 > use fallback at restart = false
 > running db nodes   = 
 > ['','','','','','','','','','','']
 > stopped db nodes   = ['','']
 > master node tables = []
 > remote             = [cpl_script_graph,
 >                       eventdata,
 >                       forward,
 >                       gruu,
 >                       numbers,
 >                       phone,
 >                       regexproute,
 >                       user]
 > ram_copies         = [call,schema]
 > disc_copies        = []
 > disc_only_copies   = []
 > [{'',ram_copies},
 > {'',ram_copies}] = [call]
 > [{'',ram_copies},
 >  {'',ram_copies},
 >  {'',ram_copies},
 >  {'',ram_copies},
 >  {'',disc_copies},
 >  {'',disc_copies},
 >  {'',ram_copies},
 >  {'',ram_copies},
 >  {'',ram_copies},
 >  {'',ram_copies},
 >  {'',ram_copies}] = [schema]
 > [{'',ram_copies},
 >  {'',ram_copies},
 >  {'',disc_copies},
 >  {'',disc_copies}] = [eventdata]
 > [{'',disc_copies},
 >  {'',disc_copies}] = [gruu,
 >                                               cpl_script_graph,
 >                                               regexproute,
 >                                               phone,
 >                                               forward,
 >                                               numbers,
 >                                               user]
 > 759 transactions committed, 0 aborted, 0 restarted, 0 logged to disc
 > 0 held locks, 0 in queue; 0 local transactions, 0 remote
 > 0 transactions waits for other nodes: []
 > ok
 > ()59>                          
 > 
 > /Fredrik
 > _______________________________________________
 > erlang-questions mailing list
 > 
 > http://www.erlang.org/mailman/listinfo/erlang-questions




More information about the erlang-questions mailing list