Comment OE Replication

some details:::::
OE 11.4 moving to 11.7.6 soon
HP-UX 11.31 migrating to RH Linux in the future
There are of many types of replication, why can't we just use Zerto or hardware block level replication, cause Progress does not support/certify any other products.
Cons: Cost
Crashes enough times to make you not want to buy OE repl.
Can't restore Target.
 
Thanks for the details, however I'm still left with questions.

What is your current data replication strategy?
What are your current and desired RPO/RTO?
What other valid data replication strategies are you considering? (Pro/con requires a comparison with something.)

Server/hypervisor/SAN-level replication solutions aren't guaranteed to work with an OpenEdge database because they don't guarantee the correct ordering of block updates, and because part of the state of a running database is the modified buffers in database shared memory, which file- or block-based tools do not know how to replicate. So a copy of a database made with such a tool is likely to be logically inconsistent and unusable.

You can use such tools if you first quiesce the database (and ensure the quiet point has begun), and if you ensure that you understand the structure of the database and you copy all extents.

dbacdog said:
Crashes enough times to make you not want to buy OE repl.
Click to expand...
What crashes? Your database? If so, your current environment may have some other issue that should be explored in a separate thread. That isn't normal.

dbacdog said:
Can't restore Target.
Click to expand...
I have no idea what this means. There is a limit to how much help people can provide for you if you aren't going to put in the effort to post coherent thoughts in complete sentences.
 
sorry - just busy
You are right, i should put in the effort.
I get the notifications and alerts from Progress, and some of the feeds are OE replication target crashes after minor upgrade from 11.6 to 11.7 or Target OpenEdge database on replication went down with errors, etc..... I just have some concerns on how stable is OE repl?

Our DR solution as of now is Netbackup replication with AI replication (roll forward).
Our RPO is 24hrs
 
Those seem pretty large, though of course it depends on your change note volume. How many extent switches do you average per day?

I would guess your present scheme already more than meets your stated needs. If your RPO is 24 hours, and that genuinely meets the business' needs, then what is the business driver for spending money on OE Replication? Even if it went down into the 5-15 minute range, you could still manage that with AI log shipping. RPO of 24 hours definitely doesn't require OE Repl.
 
dbacdog said:
We do between 150 to 170 extent switches per day.
Click to expand...
So, about one every 9 minutes on average. If you have a process to copy them over to the DR server and apply them to a warm spare DB, and you're monitoring it on an ongoing basis to ensure it's working, your actual RPO is already quite a bit better than your target.

Pro:
  • OE Repl would further reduce your RPO.
Con:
  • There is currently no business case for doing so.
(Not a complete list of either, but should suffice for now.)
 
A 24 hour RPO might be making sense to someone whose head is still stuck in 1990 but there are not many businesses in 2020 that will find that palatable.

The business will not give two hoots that they have supposedly agreed to that RPO if an unfortunate event should ever lead to you being tempted to make such a claim.
 
Just because there are reports about OE Replication in PANS doesn't mean it's unstable or unusable. By that logic I should avoid using OpenEdge at all because there are problems reported in PANS every day. In general it just works, so long as you monitor everything. I have NEVER had any avoidable problems with it.
 
OpenEdge Replication can be a PITA to setup and manage and like all software it has some bugs but it works just fine.

The “replication sets” (11.6?) enhancements are actually kind of nice.
 
I know Tom, short story is, year ago we had nada, from nothing to 24 hours is not that great, however, it beats 30 days. The business only gave us a few dollars, so we have so crappy storage and servers at our DR site. But now they are willing to spend money, so we are on DC 2.0.
 
You must log in or register to reply here.
Back
Top