Opened 14 years ago
Last modified 11 years ago
#519 reopened defect
huge database cannot be cut down again — at Initial Version
| Reported by: | anonymous | Owned by: | |
|---|---|---|---|
| Priority: | minor | Milestone: | 0.7 |
| Component: | Administration interface | Version: | dev |
| Keywords: | Cc: | mpotter@… | |
| Operating System: |
Description
I use the svn version of bitten with Apache 2.4 and trac 0.10.4 on a Windows Server 2003 RC2 SP 1.
Over month many builds have produced a huge database of 1,85 GB. The build system runs about 10 platforms in parallel. We have committed about 3000 revisions over the past 12 months. Each build consists of 5 steps, each one with a complete command-line output of the compiler+linker we run.
It seems trac stores all the logging of the builds in its project database and as we reached a size nearly of 2 GB, the whole server system becomes slow and even Apache crashes from time to time.
I want to cut down the database and free it from the ancient build logging, but I've found no way to do so. The database size remains the same although I set a new (much higher) start revision in the admin web page of the bitten build configuration. I expected the database to reduce its size but it has not happen.
Don't know what to do now. Seems it's not supported (?)
