Files
aptly/api/db.go
Ryan Gonzalez 19a705f80d Split reflists to share their contents across snapshots
In current aptly, each repository and snapshot has its own reflist in
the database. This brings a few problems with it:

- Given a sufficiently large repositories and snapshots, these lists can
  get enormous, reaching >1MB. This is a problem for LevelDB's overall
  performance, as it tends to prefer values around the confiruged block
  size (defaults to just 4KiB).
- When you take these large repositories and snapshot them, you have a
  full, new copy of the reflist, even if only a few packages changed.
  This means that having a lot of snapshots with a few changes causes
  the database to basically be full of largely duplicate reflists.
- All the duplication also means that many of the same refs are being
  loaded repeatedly, which can cause some slowdown but, more notably,
  eats up huge amounts of memory.
- Adding on more and more new repositories and snapshots will cause the
  time and memory spent on things like cleanup and publishing to grow
  roughly linearly.

At the core, there are two problems here:

- Reflists get very big because there are just a lot of packages.
- Different reflists can tend to duplicate much of the same contents.

*Split reflists* aim at solving this by separating reflists into 64
*buckets*. Package refs are sorted into individual buckets according to
the following system:

- Take the first 3 letters of the package name, after dropping a `lib`
  prefix. (Using only the first 3 letters will cause packages with
  similar prefixes to end up in the same bucket, under the assumption
  that packages with similar names tend to be updated together.)
- Take the 64-bit xxhash of these letters. (xxhash was chosen because it
  relatively good distribution across the individual bits, which is
  important for the next step.)
- Use the first 6 bits of the hash (range [0:63]) as an index into the
  buckets.

Once refs are placed in buckets, a sha256 digest of all the refs in the
bucket is taken. These buckets are then stored in the database, split
into roughly block-sized segments, and all the repositories and
snapshots simply store an array of bucket digests.

This approach means that *repositories and snapshots can share their
reflist buckets*. If a snapshot is taken of a repository, it will have
the same contents, so its split reflist will point to the same buckets
as the base repository, and only one copy of each bucket is stored in
the database. When some packages in the repository change, only the
buckets containing those packages will be modified; all the other
buckets will remain unchanged, and thus their contents will still be
shared. Later on, when these reflists are loaded, each bucket is only
loaded once, short-cutting loaded many megabytes of data. In effect,
split reflists are essentially copy-on-write, with only the changed
buckets stored individually.

Changing the disk format means that a migration needs to take place, so
that task is moved into the database cleanup step, which will migrate
reflists over to split reflists, as well as delete any unused reflist
buckets.

All the reflist tests are also changed to additionally test out split
reflists; although the internal logic is all shared (since buckets are,
themselves, just normal reflists), some special additions are needed to
have native versions of the various reflist helper methods.

In our tests, we've observed the following improvements:

- Memory usage during publish and database cleanup, with
  `GOMEMLIMIT=2GiB`, goes down from ~3.2GiB (larger than the memory
  limit!) to ~0.7GiB, a decrease of ~4.5x.
- Database size decreases from 1.3GB to 367MB.

*In my local tests*, publish times had also decreased down to mere
seconds but the same effect wasn't observed on the server, with the
times staying around the same. My suspicions are that this is due to I/O
performance: my local system is an M1 MBP, which almost certainly has
much faster disk speeds than our DigitalOcean block volumes. Split
reflists include a side effect of requiring more random accesses from
reading all the buckets by their keys, so if your random I/O
performance is slower, it might cancel out the benefits. That being
said, even in that case, the memory usage and database size advantages
still persist.

Signed-off-by: Ryan Gonzalez <ryan.gonzalez@collabora.com>
2025-02-15 23:49:21 +01:00

230 lines
6.8 KiB
Go
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
package api
import (
"fmt"
"sort"
"github.com/aptly-dev/aptly/aptly"
"github.com/aptly-dev/aptly/database"
"github.com/aptly-dev/aptly/deb"
"github.com/aptly-dev/aptly/task"
"github.com/aptly-dev/aptly/utils"
"github.com/gin-gonic/gin"
)
// @Summary DB Cleanup
// @Description **Cleanup Aptly DB**
// @Description Database cleanup removes information about unreferenced packages and deletes files in the package pool that arent used by packages anymore.
// @Description It is a good idea to run this command after massive deletion of mirrors, snapshots or local repos.
// @Tags Database
// @Produce json
// @Param _async query bool false "Run in background and return task object"
// @Success 200 {object} string "Output"
// @Failure 404 {object} Error "Not Found"
// @Router /api/db/cleanup [post]
func apiDbCleanup(c *gin.Context) {
resources := []string{string(task.AllResourcesKey)}
maybeRunTaskInBackground(c, "Clean up db", resources, func(out aptly.Progress, detail *task.Detail) (*task.ProcessReturnValue, error) {
var err error
collectionFactory := context.NewCollectionFactory()
// collect information about referenced packages and their reflist buckets...
existingPackageRefs := deb.NewSplitRefList()
existingBuckets := deb.NewRefListDigestSet()
reflistMigration := collectionFactory.RefListCollection().NewMigration()
out.Printf("Loading mirrors, local repos, snapshots and published repos...")
err = collectionFactory.RemoteRepoCollection().ForEach(func(repo *deb.RemoteRepo) error {
sl := deb.NewSplitRefList()
e := collectionFactory.RefListCollection().LoadCompleteAndMigrate(sl, repo.RefKey(), reflistMigration)
if e != nil && e != database.ErrNotFound {
return e
}
existingPackageRefs = existingPackageRefs.Merge(sl, false, true)
existingBuckets.AddAllInRefList(sl)
return nil
})
if err != nil {
return nil, err
}
err = collectionFactory.LocalRepoCollection().ForEach(func(repo *deb.LocalRepo) error {
sl := deb.NewSplitRefList()
e := collectionFactory.RefListCollection().LoadCompleteAndMigrate(sl, repo.RefKey(), reflistMigration)
if e != nil && e != database.ErrNotFound {
return e
}
existingPackageRefs = existingPackageRefs.Merge(sl, false, true)
existingBuckets.AddAllInRefList(sl)
return nil
})
if err != nil {
return nil, err
}
err = collectionFactory.SnapshotCollection().ForEach(func(snapshot *deb.Snapshot) error {
sl := deb.NewSplitRefList()
e := collectionFactory.RefListCollection().LoadCompleteAndMigrate(sl, snapshot.RefKey(), reflistMigration)
if e != nil {
return e
}
existingPackageRefs = existingPackageRefs.Merge(sl, false, true)
existingBuckets.AddAllInRefList(sl)
return nil
})
if err != nil {
return nil, err
}
err = collectionFactory.PublishedRepoCollection().ForEach(func(published *deb.PublishedRepo) error {
if published.SourceKind != deb.SourceLocalRepo {
return nil
}
for _, component := range published.Components() {
sl := deb.NewSplitRefList()
e := collectionFactory.RefListCollection().LoadCompleteAndMigrate(sl, published.RefKey(component), reflistMigration)
if e != nil {
return e
}
existingPackageRefs = existingPackageRefs.Merge(sl, false, true)
existingBuckets.AddAllInRefList(sl)
}
return nil
})
if err != nil {
return nil, err
}
err = reflistMigration.Flush()
if err != nil {
return nil, err
}
if stats := reflistMigration.Stats(); stats.Reflists > 0 {
out.Printf("Split %d reflist(s) into %d bucket(s) (%d segment(s))",
stats.Reflists, stats.Buckets, stats.Segments)
}
// ... and compare it to the list of all packages
out.Printf("Loading list of all packages...")
allPackageRefs := collectionFactory.PackageCollection().AllPackageRefs()
toDelete := allPackageRefs.Subtract(existingPackageRefs.Flatten())
// delete packages that are no longer referenced
out.Printf("Deleting unreferenced packages (%d)...", toDelete.Len())
// database can't err as collection factory already constructed
db, _ := context.Database()
if toDelete.Len() > 0 {
batch := db.CreateBatch()
toDelete.ForEach(func(ref []byte) error {
collectionFactory.PackageCollection().DeleteByKey(ref, batch)
return nil
})
err = batch.Write()
if err != nil {
return nil, fmt.Errorf("unable to write to DB: %s", err)
}
}
bucketsToDelete, err := collectionFactory.RefListCollection().AllBucketDigests()
if err != nil {
return nil, err
}
bucketsToDelete.RemoveAll(existingBuckets)
out.Printf("Deleting unreferenced reflist buckets (%d)...", bucketsToDelete.Len())
if bucketsToDelete.Len() > 0 {
batch := db.CreateBatch()
err := bucketsToDelete.ForEach(func(digest []byte) error {
return collectionFactory.RefListCollection().UnsafeDropBucket(digest, batch)
})
if err != nil {
return nil, err
}
if err := batch.Write(); err != nil {
return nil, err
}
}
// now, build a list of files that should be present in Repository (package pool)
out.Printf("Building list of files referenced by packages...")
referencedFiles := make([]string, 0, existingPackageRefs.Len())
err = existingPackageRefs.ForEach(func(key []byte) error {
pkg, err2 := collectionFactory.PackageCollection().ByKey(key)
if err2 != nil {
tail := ""
return fmt.Errorf("unable to load package %s: %s%s", string(key), err2, tail)
}
paths, err2 := pkg.FilepathList(context.PackagePool())
if err2 != nil {
return err2
}
referencedFiles = append(referencedFiles, paths...)
return nil
})
if err != nil {
return nil, err
}
sort.Strings(referencedFiles)
// build a list of files in the package pool
out.Printf("Building list of files in package pool...")
existingFiles, err := context.PackagePool().FilepathList(out)
if err != nil {
return nil, fmt.Errorf("unable to collect file paths: %s", err)
}
// find files which are in the pool but not referenced by packages
filesToDelete := utils.StrSlicesSubstract(existingFiles, referencedFiles)
// delete files that are no longer referenced
out.Printf("Deleting unreferenced files (%d)...", len(filesToDelete))
countFilesToDelete := len(filesToDelete)
taskDetail := struct {
TotalNumberOfPackagesToDelete int
RemainingNumberOfPackagesToDelete int
}{
countFilesToDelete, countFilesToDelete,
}
detail.Store(taskDetail)
if countFilesToDelete > 0 {
var size, totalSize int64
for _, file := range filesToDelete {
size, err = context.PackagePool().Remove(file)
if err != nil {
return nil, err
}
taskDetail.RemainingNumberOfPackagesToDelete--
detail.Store(taskDetail)
totalSize += size
}
out.Printf("Disk space freed: %s...", utils.HumanBytes(totalSize))
}
out.Printf("Compacting database...")
return nil, db.CompactDB()
})
}