On 10-06-2012 02:05:54 -0700, Majid Azimi wrote:
Hi guys,
What types of clustering does monetdb support? Sharding? Master/Slave?
If you want it out-of-the-box, none.
if it is Master/Slave does it support multi master? or all writes must go to master node?
If you shard your data, you'll have to direct the writes to the appropriate nodes yourself.
In clustering documentation (http://www.monetdb.org/Documentation/Cookbooks/Recipes/Cluster) there is paragraph:
willem% monetdb -h cruquius -P mypassphrase set shared=pump/2/harvey pump willem% monetdb -h lijnden -P mypassphrase set shared=pump/2/fox pump willem% monetdb -h leeghwater -P mypassphrase set shared=pump/1/harvey pump
What are those 1, 2, fox, harvey?
Fox and Harvey [1] (from Hayle) are part of the names we give to the three "pumps", based on the manufacturer of their Cornwall steam engine. (Balances and valves Vlissingen/Dudok van Heel (Werkspoor/Stork) [2].) As you can see, Leeghwater and Cruquius came from Harvey & Co., only Lijnden from Fox & Co. This is because both the engines for Cruquius and Lijnden had to be produced at the same time, and Harvey couldn't do that. Leeghwater was produced before as semi-successful try-out, by Harvey. Now in our example we assign the tag "pump/X/name" to the three databases. This means we can match any pump with the pattern "pump/*" (or "pump/*/*"). However, we might only want to match the engines from Harvey, for which we could use the patterns "pump/*/harvey", "*/*/harvey" or "*/harvey", depending on the required semantics. Likewise, we could want to select only the two 2nd generation pumps (after the first try-out) with patterns "*/2/*" and "pump/2/*". See also REMOTE DATABASES section in monetdbd(1) [3]. [1] http://en.wikipedia.org/wiki/Harvey_%26_Co [2] http://nl.wikipedia.org/wiki/Werkspoor [3] http://www.monetdb.org/Documentation/monetdbd-man-page