Ta yaya ake zaɓar adadin shards da replicas na Elasticsearch?

Question

IT Interview Practice · Accepted Answer

Zaɓi primary shards don raba data da aiki; zaɓi replicas don ƙarin kwafi da ƙarfin bincike. Fara daga workload da buƙatar farfaɗowa, ba adadin shards guda da ya dace da kowa ba.

## Lissafa hasashe na farko

Ka ɗauka primary index data ita ce 120 GB kuma gwaji ya nuna kusan 30 GB ga kowane primary shard ya dace da workload ɗinka:

```text
120 GB / 30 GB = 4 primary shards
1 replica per primary = 4 additional shard copies
Total = 8 shard copies, roughly 240 GB of shard data
```

Waɗannan inputs na ƙiyasin girma na misali ne. Hasashen storage bai haɗa da girma nan gaba da operating space na ayyuka kamar segment merges, waɗanda ke haɗa fayilolin index a bango, ba.

## Me ya sa ba a ci gaba da ƙara shards kawai ba?

Ƙarin shards na iya raba aiki a nodes, amma kowanne yana ƙara management overhead kuma bincike zai iya bazuwa zuwa wurare da yawa. Manyan shards suna rage overhead amma suna iya ɗaukar lokaci wajen farfaɗowa ko motsawa. Replicas suna ƙara redundancy amma kuma storage da aikin rubutu; ba backups ba ne.

## Gwada yanayin matsala ma

Auna peak reads da writes tare, sannan gwada rasa node. Sauran nodes suna buƙatar isasshen disk, memory da ƙarfin aiki don farfaɗowa cikin lokacin da ka sa. Ga time-based data mai girma, rollover yana ƙirƙirar sabon index idan an kai sharadin da aka zaɓa, yana taimakawa wajen iyakance girman index. Waɗannan shawarwari sun shafi deployments da mai amfani ke sarrafa shard settings.

[Takardun hukuma](https://www.elastic.co/docs/deploy-manage/production-guidance/optimize-performance/size-shards).

Lissafa hasashe na farko

Me ya sa ba a ci gaba da ƙara shards kawai ba?

Gwada yanayin matsala ma

Tambayoyi masu alaƙa