Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
/ 2004-08-19 16:36:11 +0100 \ German Gutierrez: > Hi nettiquette says don't reply right into some thread with a new topic. thanks. > I chkconfig --add drbd to add service to startup . > > But , when i start Node 1 , If the second is not > available , Node 1 hang waiting for Node 2 on startup > , So ... there might be a reason for it... we want to be in connected state before some cluster manager takes control... with drb 0.7 and a sane cluster manager, this is no longer strictly needed, but will reduce the likelihood of data loss in some strange, but not too far fetched, failure scenarios. its not only about high availability, its also about not losing data. and the behavious is even configurable... read the documentation. I admit it is not an O'Reilly book (maybe I one day write one; "High Availability of your Data Storage with DBRD and linux-ha" might be a nice working title...), but at least it is available. even if you only have a look at the example drbd.conf file, there should be a comment in the "startup" section... > What are your recommendations for handling drbd > Service ? > > Is possible add drbd like resouce ? What is the best > way ? possible, yes. not intended. stongly advised against in typical deployments. the "resource" (in the heartbeat sense) is drbddisk. hth, Lars Ellenberg -- please use the "List-Reply" function of your email client.