Cyrus-IMAP uses Xapian, too. It feeds new email into a Xapian database. Over time, the Xapian database gets pretty big and each change tooks longer and longer. The recommendation is to do regularly “database compaction (repack)”. The old data is compressed into a smaller database, new data goes into a recently created database. The recently created database is small, thus adding data to it and saving the changes is also fast. On search, both the new and old, compacted databases are used.
In the documentation of xapian-haystack I do not find any text about this compaction and having many database files (tiers), used simultaneously when searching.
Cyrus-IMAP uses Xapian, too. It feeds new email into a Xapian database. Over time, the Xapian database gets pretty big and each change tooks longer and longer. The recommendation is to do regularly “database compaction (repack)”. The old data is compressed into a smaller database, new data goes into a recently created database. The recently created database is small, thus adding data to it and saving the changes is also fast. On search, both the new and old, compacted databases are used.
In the documentation of xapian-haystack I do not find any text about this compaction and having many database files (tiers), used simultaneously when searching.
https://lists.tartarus.org/pipermail/xapian-discuss/2014-October/009112.html