Enterprise Replication error when attempting to delete a replicate

Hello, When attempting to delete a replicate, the following error is raised: $ cdr del repl r_bln_db_b0400_db_sk57command failed -- Attempt to perform invalid operation including shadow replicates (122) Additional info: ----- onstat - IBM Informix Dynamic Server Version 10.00.FC4 ----- The server is an AIX 5.3 ML04. ----- No error message in online.log. No error message from "cdr error ". There are no locks on the replicated table (source table/destination table). ----- $ cdr list repl r_bln_db_b0400_db_sk57 DEFINED REPLICATES ATTRIBUTES ------------------------------ REPLICATE: r_bln_db_b0400_db_sk57 STATE: Inactive ON:g_sk57 CONFLICT: Ignore FREQUENCY: immediate QUEUE SIZE: 0 PARTICIPANT: b0400_db:ibanks.bln_db OPTIONS: transaction,ris,ats REPLID: 33227027 / 0x1fb0113 REPLMODE: PRIMARY ON:g_sk57 APPLY-AS: INFORMIX ON:g_sk57 REPLTYPE: Master cdr list server SERVER ID STATE STATUS QUEUE CONNECTION CHANGED ----------------------------------------------------------------------- g_sk57 507 Active Local 0 g_sk67 607 Active Connected 1728 Aug 1 13:19:10 Repeating the cdr list server after attempting to delete the replicate again: cdr list server SERVER ID STATE STATUS QUEUE CONNECTION CHANGED ----------------------------------------------------------------------- g_sk57 507 Active Local 0 g_sk67 607 Active Connected 1860 Aug 1 13:19:10 Checking the network threads on both servers (onstat -g ntt | grep -i cdr): 70000005135aa70 CDRNsT607 4884 13:19:10 13:35:10 13:35:11 700000052342d20 cdraccept 1101 13:19:11 13:35:11 13:35:11 cdr list server SERVER ID STATE STATUS QUEUE CONNECTION CHANGED ----------------------------------------------------------------------- g_sk57 507 Active Local 0 g_sk67 607 Active Connected 2256 Aug 1 13:19:10 Again we can see the increased queue size for g_sk67, even if it reports communication on the cdr network threads. Any ideas? Regards, Bogdan BOTEZ.

Bogdan Botez wrote: > Hello, > > When attempting to delete a replicate, the following error is raised: > > $ cdr del repl r_bln_db_b0400_db_sk57> command failed -- Attempt to perform invalid operation including > shadow replicates > (122) The shadow replicate should eventually automatically delete itself, but only when it is sure that it is no longer needed. This might take some time. However, you can manually delete shadow replicates. 1) first do "cdr list repl" 2) Then look for the shadow replicates based on r_bln_db_b0400_db_sk57 3) Then issue "cdr delete repl" for that replicate. As an example, I have a replicate "repl1" and a shadow replicate based on repl1. The shadow replicate has a name of repl1a. The output of cdr list repl looks like .... $ i cdr list repl CURRENTLY DEFINED REPLICATES ------------------------------- REPLICATE: repl1 STATE: Active ON:cdr1 CONFLICT: Timestamp FREQUENCY: immediate QUEUE SIZE: 0 PARTICIPANT: cmpdb:mpruet.cmptab1 OPTIONS: transaction,ris,ats,triggers,fullrow REPLID: 65537 / 0x10001 REPLMODE: PRIMARY ON:cdr1 APPLY-AS: INFORMIX ON:cdr1 REPLICATE: repl1a STATE: Active ON:cdr1 CONFLICT: Timestamp FREQUENCY: immediate QUEUE SIZE: 0 PARTICIPANT: cmpdb:mpruet.cmptab1 OPTIONS: row,fullrow REPLID: 65539 / 0x10003 REPLMODE: PRIMARY ON:cdr1 APPLY-AS: INFORMIX ON:cdr1 REPLTYPE: Shadow PARENT REPLICATE: repl1 So I could issue the following: cdr delete repl repl1a cdr delete repl repl1 > > Additional info: > ----- > onstat - > IBM Informix Dynamic Server Version 10.00.FC4 > ----- > The server is an AIX 5.3 ML04. > ----- > No error message in online.log. > No error message from "cdr error ". > There are no locks on the replicated table (source table/destination > table). > ----- > $ cdr list repl r_bln_db_b0400_db_sk57> > DEFINED REPLICATES ATTRIBUTES > ------------------------------ > REPLICATE: r_bln_db_b0400_db_sk57 > STATE: Inactive ON:g_sk57 > CONFLICT: Ignore > FREQUENCY: immediate > QUEUE SIZE: 0 > PARTICIPANT: b0400_db:ibanks.bln_db > OPTIONS: transaction,ris,ats > REPLID: 33227027 / 0x1fb0113 > REPLMODE: PRIMARY ON:g_sk57 > APPLY-AS: INFORMIX ON:g_sk57 > REPLTYPE: Master > > cdr list server > SERVER ID STATE STATUS QUEUE CONNECTION CHANGED > ----------------------------------------------------------------------- > g_sk57 507 Active Local 0 > g_sk67 607 Active Connected 1728 Aug 1 13:19:10 > > Repeating the cdr list server after attempting to delete the replicate > again: > > cdr list server > SERVER ID STATE STATUS QUEUE CONNECTION CHANGED > ----------------------------------------------------------------------- > g_sk57 507 Active Local 0 > g_sk67 607 Active Connected 1860 Aug 1 13:19:10 > > Checking the network threads on both servers (onstat -g ntt | grep -i cdr): > > 70000005135aa70 CDRNsT607 4884 13:19:10 13:35:10 13:35:11 > > 700000052342d20 cdraccept 1101 13:19:11 13:35:11 13:35:11 > > cdr list server > SERVER ID STATE STATUS QUEUE CONNECTION CHANGED > ----------------------------------------------------------------------- > g_sk57 507 Active Local 0 > g_sk67 607 Active Connected 2256 Aug 1 13:19:10 > > Again we can see the increased queue size for g_sk67, even if it > reports communication on the cdr network threads. > > Any ideas? > > Regards, > Bogdan BOTEZ.

Thanks for your answers. If there is another opinion, I still want to hear about it... Regards, Bogdan BOTEZ.