It seems to me that using multiple controllers automatically supports having faul-tolerancy in recovery logs. So if you have 2 controllers, you should have 2 recovery logs for fault-tolerance. The recovery log configuration is part of each controller's config, so each controller is accessing its own recoverylog db (in a fault-tolerant situation). It seems it's common for these dbs to sit on each controller's local machine, although this isn't required. As long as you make <RequestManager> <RecoveryLog> config block in each controller, those controllers will have access to those recoverylog databases (which all should have the same information in them (in RAIDb-1 anyway)). So if on a controller you had the config block: <RecoveryLog driver="org.hsqldb.jdbcDriver" url="jdbc:hsqldb:hsql://localhost:9003"> </RecoveryLog> Then you're telling that controller the recoverylog will be in a HSQL type database on port 9003 of the localhost. Note, that this is what *I* get from the documentation anyway, as I still haven't downloaded and installed sequoia yet.
-Tony --------------------------- Manager, IT Operations Format Dynamics, Inc. 303-573-1800x27 [EMAIL PROTECTED] http://www.formatdynamics.com -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of [EMAIL PROTECTED] Sent: Wednesday, April 09, 2008 7:04 AM To: Sequoia general mailing list Subject: Re: [Sequoia] Fault tolerance configuration question Emmanuel Cecchet wrote: > Jamin wrote: >> If I'm understanding correctly for the RecoveryLog to be fault >> tolerant it would need to be pointed to yet another virtual database >> (V2) which would need to be on at least two servers. Assuming I put >> it on M1 and M2, then when either M1 or M2 fail the RecoveryLog would >> become degraded and need to be recovered just like the original >> virtual database (V1) would it not? > > This is not really useful since as soon as you lose the controller, the > recovery log becomes useless. Yet this is what the Sequoia User's Guide talks about for a fault tolerant Recovery Log. From section 7.5.3, http://sequoia.continuent.org/doc/2.10/userGuide/ar01s07.html#N10809: "it is possible to make the recovery log fault tolerant by redirecting it to a Sequoia controller (even self) that will distribute and replicate the log content on several backends." However your previous statement: > This is why you need to replicate the controller (and as there > is one recovery log per virtual database per controller, you will also > replicate the recovery log). seems to indicate that simply having multiple Sequoia controllers will automatically provide fault tolerance of the Recovery Log. Where should these Recovery Log databases be stored? If they are stored on the MySQL backends I still see them running the same problem when one of the backends fails. Are you instead suggesting that the Recovery Log be stored on the controller? If so, would this result in each controller having its own independent copy of the Recovery Log? Also, is there a suggested backend to use for the storage? I'm sorry if it seems I'm missing something obvious, but the documentation really doesn't make this very clear. In fact it seems to contradict what you're saying (as indicated above). >> If this is covered in the documentation, perhaps someone could point >> me to what I've missed? > The documentation assumes upfront that you are replicating the > controllers if you want availability. > This is probably hidden somewhere in the basic concepts part like in > http://sequoia.continuent.org/doc/infocenter/index.jsp?topic=/org.contin uent.sequoia.doc/html/Sequoia_URL_options.html That documentation link appears to be for 3.x, whereas the latest stable release (according to the website) is 2.10.9 which is what I've been trying to use. -- Jamin W. Collins _______________________________________________ Sequoia mailing list [email protected] https://forge.continuent.org/mailman/listinfo/sequoia _______________________________________________ Sequoia mailing list [email protected] https://forge.continuent.org/mailman/listinfo/sequoia
