fork of https://github.com/sourcegraph/zoekt
Add ShardPrefixOverride to support custom shard naming (#1005)
Co-authored-by: Ravi Kumar <rkumar@gitlab.com>
author
Ravi Kumar
co-author
Ravi Kumar
committer
GitHub
date
(Mar 18, 2026, 2:22 PM +0200)
commit
6b56e687
6b56e6873c5f71e161cb57b88bf79c0b096cbb6d
parent
c747a3bc
c747a3bccc2a4a427204ac08eea62a522df6d2ec
+17
-7
+17
-7
index/builder.go
+17
-7
index/builder.go
···
61
61
// IndexDir is a directory that holds *.zoekt index files.
62
62
IndexDir string
63
63
64
+
// ShardPrefixOverride sets the prefix for shards name
65
+
ShardPrefixOverride string
66
+
64
67
// SizeMax is the maximum file size
65
68
SizeMax int
66
69
···
182
185
fs.IntVar(&o.ShardMax, "shard_limit", x.ShardMax, "maximum corpus size for a shard")
183
186
fs.IntVar(&o.Parallelism, "parallelism", x.Parallelism, "maximum number of parallel indexing processes.")
184
187
fs.StringVar(&o.IndexDir, "index", x.IndexDir, "directory for search indices")
188
+
fs.StringVar(&o.ShardPrefixOverride, "shard_prefix_override", x.ShardPrefixOverride, "prefix for shard name")
185
189
fs.BoolVar(&o.CTagsMustSucceed, "require_ctags", x.CTagsMustSucceed, "If set, ctags calls must succeed.")
186
190
fs.Var(largeFilesFlag{o}, "large_file", "A glob pattern where matching files are to be index regardless of their size. You can add multiple patterns by setting this more than once.")
187
191
···
214
218
args = append(args, "-index", o.IndexDir)
215
219
}
216
220
221
+
if o.ShardPrefixOverride != "" {
222
+
args = append(args, "-shard_prefix_override", o.ShardPrefixOverride)
223
+
}
224
+
217
225
if o.CTagsMustSucceed {
218
226
args = append(args, "-require_ctags")
219
227
}
···
334
342
}
335
343
336
344
func (o *Options) shardNameVersion(version, n int) string {
337
-
var prefix string
345
+
prefix := o.ShardPrefixOverride // ShardPrefixOverride takes precedence to support custom shard naming strategies
338
346
339
-
// Sourcegraph specific: We use IDs in shard names on multi-tenant
340
-
// instances to prevent conflicts.
341
-
if tenant.UseIDBasedShardNames() {
342
-
prefix = fmt.Sprintf("%09d_%09d", o.RepositoryDescription.TenantID, o.RepositoryDescription.ID)
343
-
} else {
344
-
prefix = o.RepositoryDescription.Name
347
+
if prefix == "" {
348
+
// Sourcegraph specific: We use IDs in shard names on multi-tenant
349
+
// instances to prevent conflicts.
350
+
if tenant.UseIDBasedShardNames() {
351
+
prefix = fmt.Sprintf("%09d_%09d", o.RepositoryDescription.TenantID, o.RepositoryDescription.ID)
352
+
} else {
353
+
prefix = o.RepositoryDescription.Name
354
+
}
345
355
}
346
356
347
357
return shardName(o.IndexDir, prefix, version, n)