[Remote Store] Avoiding Translog Deletion in case of relocating shards

[gbbafna]

We need to address #9603 holistically. Whenever there are relocating primary shards, we need to skip deletion during that time period.

Temporary workarounds :

1. [Remote Store] Add extra buffer before deleting older generations of translog #10817
2. Retry download of RemoteFSTranslog to fix transient race conditions #9565

Alternatives

#9603

[Bukhtawar]

Lets add a proposal for the same.

[sachinpkale]

[Storage Triage - attendees 1 2 3 4 5 6 7 8 9 10 ]

@gbbafna Please add proposal for the mitigation/solution of the issue