Remove obsolete TSM1 max points per block

[andyfeller]

Originally posted on mailing list.

When the TSM1 compaction process was updated in commit 756421ec4a5aeea994603175e73941e461c2b69a, the implementation used the default max points per block constant rather than the configured value for those who specify it:

 11 const (  
 ..
 36     // DefaultMaxPointsPerBlock is the maximum number of points in an encoded
 37     // block in a TSM file
 38     DefaultMaxPointsPerBlock = 1000
 ..
 42 )
 43 
 44 // Config holds the configuration for the tsbd package.
 45 type Config struct { 
 ..
 61     MaxPointsPerBlock              int           `toml:"max-points-per-block"`

There are 5 places affected in HEAD on master:

 278 // Plan returns a set of TSM files to rewrite for level 4 or higher.  The planning returns
 279 // multiple groups if possible to allow compactions to run concurrently.
 280 func (c *DefaultPlanner) Plan(lastWrite time.Time) []CompactionGroup {
 281     generations := c.findGenerations()
 282 
 283     // first check if we should be doing a full compaction because nothing has been written in a long time
 284     if c.CompactFullWriteColdDuration > 0 && time.Now().Sub(lastWrite) > c.CompactFullWriteColdDuration && len(generations) > 1 {
 285         var tsmFiles []string 
 286         for i, group := range generations {
 287             var skip bool
 288 
 289             // Skip the file if it's over the max size and contains a full block and it does not have any tombstones
 290             if group.size() > uint64(maxTSMFileSize) && c.FileStore.BlockCount(group.files[0].Path, 1) == tsdb.DefaultMaxPointsPerBlock && !group.hasTombstones() {
 291                 skip = true
 292             }
 ...
 357         // Skip the file if it's over the max size and contains a full block or the generation is split
 358         // over multiple files.  In the latter case, that would mean the data in the file spilled over
 359         // the 2GB limit.
 360         if g.size() > uint64(maxTSMFileSize) && c.FileStore.BlockCount(g.files[0].Path, 1) == tsdb.DefaultMaxPointsPerBlock || g.count() > 1 {
 361             start = i + 1
 362         }
 ...
 403             // Skip the file if it's over the max size and it contains a full block
 404             if gen.size() >= uint64(maxTSMFileSize) && c.FileStore.BlockCount(gen.files[0].Path, 1) == tsdb.DefaultMaxPointsPerBlock && !gen.hasTombstones() {
 405                 startIndex++
 406                 continue
 407             }

 520 // WriteSnapshot will write a Cache snapshot to a new TSM files.
 521 func (c *Compactor) WriteSnapshot(cache *Cache) ([]string, error) {
 522     c.mu.RLock()
 523     opened := c.opened
 524     c.mu.RUnlock()
 525
 526     if !opened {
 527         return nil, errSnapshotsDisabled
 528     }
 529
 530     iter := NewCacheKeyIterator(cache, tsdb.DefaultMaxPointsPerBlock)

 545 // Compact will write multiple smaller TSM files into 1 or more larger files
 546 func (c *Compactor) compact(fast bool, tsmFiles []string) ([]string, error) {
 547     size := c.Size
 548     if size <= 0 {
 549         size = tsdb.DefaultMaxPointsPerBlock
 550     }

[jwilder]

That config option was supposed to be removed. That is a left-over from a earlier version of the tsm engine, but is no longer supported.

[jwilder]

Fixed via #7510