Hello, I read that reelection in cluster should take around 12s average. I have M40 atlas with 2 replica set shards. After primary shard crashed due to lack of memory, reelecting other primary took 3mins 7s. I had another crash and it took 2min 16s to reelect other primary. Is that normal to take so long time?
Hi @Eimantas_Puskorius, thank you for your post and welcome to the MongoDB community!
In general, it is not normal for a reelection to take that long. Would you mind opening a support case so we can look into this further? Also, do you have auto-scaling enabled on your Atlas clusters? Enabling Atlas auto-scaling can help preempt these cases and ensure database availability.
Thanks!
Frank