Re: Data Replication

pjhaveri wrote: > > Hi, > Sorry for late reply. I am using 7.11UC2. We are testing replication. > It is DHR. And asysnchronous mode. > > - I was trying to test if primary goes down, whether I can use seconadry as a > standard server and when primary is up, whether I can change seconadry-which > was standard to seconadry. > > So when REPLICATION WAS WORKING, I shutdown PRIMARY, made seconadry as > standard server. > using onmode -d standard command.And :-)) > > It could not change mode to standard mode. > > Probs : Seconadry could not change mode to ONLINE fm RECOVERY MODE !!! > > I tried to take it to offline mode and then tried to bring it up, It did not > work. It remained in the recovery mode. > > On the other side when we tried to bring up primary, we could not. Following > is the msg from msg log file. > *********************************************************************** > 18:36:09 DR: DRAUTO is 0 (Off) > > 18:36:15 INFORMIX-OnLine Initialized -- Shared Memory Initialized. > > 18:36:15 DR: Trying to connect to secondary server ... > > 18:36:16 DR: Primary server connected > > 18:36:17 DR: Server type incompatible > > 18:36:17 rshdr.c, line 3344, thread 23, proc id 3232, DR: Server type> incompati > ble. > > 18:36:17 PANIC: Attempting to bring system down > ****************************************************************************> > > > > > I tried > > ------------- Begin Original Message ------------- > From: attmail!internet!rmy.emory.edu!informix-list-owner (Juan R. Guzman) > Date: Mon Aug 12 19:20:29 +0000 1996 > Subject: Re: Data Replication > To: attmail!internet!rmy.emory.edu!informix-list > Content-Type: text > Content-Length: 741 > > pjhaveri wrote: > > > > Hi, > > > > We are trying Replication between 2 cities. I wanted some information about > > this. Anyone is using replication, what are the problems you have faced. > And> what type synchronous or asynchronous mode is preferable ? > > > > your input would be greatly appreciated! > > > > Thanks and Regards, > > Pritirj > > You didn't say which on-line your using or which tpe of DR (DHR or > CDR). I have 7.11UC1 and it has a bug. When you try to bring up the > secondary (from scheduled downtie or not) it won't find the checkpoint > in the log log and so it stays in recovery mode. > > I would be very interested in hearing your feedback on how it goes > because I'm thinkg of doing the same (DR is local now). > > thanks, Hi, When you take down the primary server and want to make the secondary server standard, it is possible. Here is how you do it. 1. Take your primary down. 2. Secondary recieves a DR Error, but stays on-line. 3. On secondary, Issue 'onmode -d primary <old primary server name>' 4. Secondary becomes primary, tries to connect to secondary, fails but becomes primary. 5. Now change the mode to standard, 'onmode -d standard'. That would work in both 7.11UC1 and 7.12UC1. But here are the problems associated with it. 1. Once you make one of the servers standard, DR is lost. And if the logical logs roll over, you would have start from scratch and redo the DR method. 2. You would have to implement the client redirection methods to goto the secondary and if you have redo the DR, possibly there would be some downtime involved. I did not know why you want to make the secondary standard. But here is my suggetion 1. If primary is going to be down short amount of time, do not do anything but bring the primary up. 2. If primary is down long, make secondary primary (as explained above) and keep it primary. This is assuming you have the clients redirected to seconadary server which you would have to do anyway. By this way whenever the primary comes up it automatically changes itself to secondary and stays as secondary. Repeat the procedure if needed. 3. The advantage of above scheme is you do not have to redirect your clients twice. One more caution. Do back up your logical logs to tape/disk. Do not make LTAPEDEV to /dev/null. Hope this will help. Zakeer Shaik