Drop the recommendation that --expected-shred-version be set by validators (#12244)

`--expected-shred-version` is another knob for users to get wrong and is
documentation that can get stale due to cluster restarts.  Turns out
it's also generally not required anymore either because:
1. The cluster entrypoint can always be expected to be using the correct
   shred version, and that shred version will be adopted by the new node
   (earlier this was not the case when the `solana-gossip spy` node on
   mainnet-beta.solana.com:8001 ran with shred version 0)
2. On a cluster restart, `--expected-bank-hash` is a much stronger
   assertion that the validator is starting from the correct place (and
   didn't exist when `--expected-shred-version` was first recommended)

(cherry picked from commit 4ada4d43f2)

Co-authored-by: Michael Vines <mvines@gmail.com>

docs/src/running-validator/restart-cluster.md

@@ -21,7 +21,6 @@ call this NEW\_SHRED\_VERSION and NEW\_BANK\_HASH respectively.
 Adjust your validator's arugments:
 
 ```bash
- --expected-shred-version NEW_SHRED_VERSION
  --wait-for-supermajority SLOT_X
  --expected-bank-hash NEW_BANK_HASH
 ```
@@ -51,7 +50,6 @@ Post something like the following to #announcements (adjusting the text as appro
 > ```bash
 > solana-validator
 >   --wait-for-supermajority 12961040  # <-- NEW! IMPORTANT!
->   --expected-shred-version 56096     # <-- NEW! IMPORTANT!
 >   --expected-bank-hash 6q2oTgs8FiJxra2Zo1N8tzWqo5b6uGbjmFgoWDsXxchY    # <-- NEW! IMPORTANT!
 >   --hard-fork 56096                  # <-- NEW! IMPORTANT!
 >   --entrypoint 35.203.170.30:8001    # <-- Same as before