Update - the unit test checked in as part of this fix can be used to reproduce. Just in case anyone else runs into this problem.
On Thursday, February 28, 2019 at 10:08:04 AM UTC-6, [email protected] wrote: > > I have a large number of H2 instances deployed in an embedded environment > in the field (version 1.4.194). > Certain systems have a very large number of IO's being generated. These > are mostly idle systems, very little client generated IO. The result is > burning up the mSATA devices that the DB is running on. > > The symptom is an increasing number of IO's written to the DB (MVStore, > single instance) over time. We're seeing numbers of over 600MB of data > written per hour, to a 6MB DB file, with an almost completely idle system. > The DB is used almost exclusively to store compressed text BLOB objects in > tables with just a few columns. > > Based on the Issue tagged in the subject, we think that this is the > problem cause. > > Upgrading to the latest release of 1.4.198 seems to have eliminated the > issue. > We'd like to know for the 194 release how to positively initiate the > problem as it appears sometimes and other times the DB acts normally. We > need to be able to positively reproduce the problem with the older version > in order to have confidence that the newest release fixes the issue we're > seeing. > > Thanks very much for the help, Eric. > -- You received this message because you are subscribed to the Google Groups "H2 Database" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/h2-database. For more options, visit https://groups.google.com/d/optout.
