On Jul 8, 2010, at 2:36 PM, Michael wrote: > Hello everyone; > > I am new to this list and to Couch. > > For my first real test with Couch I created a database which was updated via > another service every 5 minutes. I guess i didn't realize how much data I > was taking in, because I got distracted and came back to the server and the > couch database had grown large enough to fill all the disk space. >
one way to recover here is to use replication to clone your database to another host, then delete the original database and replicate back. if you are updating documents a lot (and not just adding new docs) replication alone will result in a much smaller file. if you aren't updating (just logging new docs) then you'll still end up with a smaller file, but maybe not quite as small. > Clearly, I can delete the database manually, recover my disk space and fix > everything, but I am learning, so bear with me. > > I seem to have found bug 808 [1] while trying to compact the database. This > makes Couchdb unresponsive. Is there an easy way to manually compact the > database, perhaps specifying a mount which has more disk space? > > [1] https://issues.apache.org/jira/browse/COUCHDB-808 > > So the next question is, how do I keep this under control? I think to keep > myself sane I will probably write a script that checks the size of the > database and if it exceeds a certain size the script will compact the > database. Is there an automatic way to do this? Anyone have any best > practices? When the wiki says "near full capacity for its write load" how do > I measure what full capacity is? if you are importing every few minutes, you are not anywhere near full write-throughput capacity, so compact whenever you feel like it. in the future we'd like to add some form of auto compaction, but it's not around yet. Have fun! Chris > > Thanks for the help. I am really enjoying Couch, it feels right, > > Michael
