Here's a smaple from the reindexing: It seems to go over the 100% mark.. I says it's an estimate, but it's far off from the estimate.
2014-12-02 13:53:27:053 INFO --> 12421.49% progress, 1,616,360,917 indexed so far (14,016 items/sec) [OIndexRebuildOutputListener] 2014-12-02 13:53:37:055 INFO --> -10428.31% progress, -1,356,996,670 indexed so far (-297,335,758 items/sec) [OIndexRebuildOutputListener] 2014-12-02 13:53:47:057 INFO --> -10427.70% progress, -1,356,916,860 indexed so far (7,981 items/sec) [OIndexRebuildOutputListener] 2014-12-02 13:53:57:057 INFO --> -10427.09% progress, -1,356,837,084 indexed so far (7,977 items/sec) [OIndexRebuildOutputListener] 2014-12-02 13:54:07:058 INFO --> -8704.30% progress, -1,132,657,911 indexed so far (22,417,917 items/sec) [OIndexRebuildOutputListener] 2014-12-02 13:54:17:059 INFO --> -8703.20% progress, -1,132,514,089 indexed so far (14,382 items/sec) [OIndexRebuildOutputListener] 2014-12-02 13:54:27:566 INFO --> -8702.21% progress, -1,132,384,614 indexed so far (12,947 items/sec) [OIndexRebuildOutputListener] 2014-12-02 13:54:37:567 INFO --> -8701.03% progress, -1,132,231,797 indexed so far (15,281 items/sec) [OIndexRebuildOutputListener] 2014-12-02 13:54:47:568 INFO --> -8699.91% progress, -1,132,086,152 indexed so far (14,564 items/sec) [OIndexRebuildOutputListener] 2014-12-02 13:54:57:569 INFO --> -8698.78% progress, -1,131,939,074 indexed so far (14,707 items/sec) [OIndexRebuildOutputListener] 2014-12-02 13:55:07:570 INFO --> -8697.58% progress, -1,131,783,540 indexed so far (15,553 items/sec) [OIndexRebuildOutputListener] 2014-12-02 13:55:17:574 INFO --> -8696.48% progress, -1,131,640,029 indexed so far (14,351 items/sec) [OIndexRebuildOutputListener] 2014-12-02 13:55:27:575 INFO --> -8695.45% progress, -1,131,505,704 indexed so far (13,432 items/sec) [OIndexRebuildOutputListener] On Tuesday, December 2, 2014 1:55:42 PM UTC-5, Simon L wrote: > > Hi there, > > We're using ODB for production uses on Amazon EC2. Our DB is over 100 GB. > > Very often, when not perfectly shutting down the database, or even > occasionally somewhat randomly, the database starts reindexing. We cannot > use the DB while it is reindexing. > > This is a _huge_ problem. In a cluster, theoretically, if a DB shuts down > and needs reindexing, I'm guessing we can remove that node from the > cluster. Practically, that is an impossible situation for many reasons. > First, whatever caused issues on the first DB will likely affect other > systems. Especially since with 1 node removed, the others will be under a > strain. I've had some experience with situations like this and it can get > very ugly very very quickly. > > > It also takes a very long time... Often, in test, we don't even allow it > to reindex because it takes hours upon hours. > > I've never seen this sort of behaviour with other DBs I've used. They've > all seemed extremely durable when it comes to things like random shutdowns. > While I cannot speak for the type of recovery-based rebuilding, even > reindexing on OrientDB seems to cause major slowdowns that are > inappropriate for production. > > > > -- --- You received this message because you are subscribed to the Google Groups "OrientDB" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
