SDSTOR-21989: check index_table existence before using it

[JacksonYao287]

if homeobject crashes after index_table is destroyed in destroy_pg and before the pg is created in baseline resync, we can not find pg index table when restarts. so we need add a sanity check to see if index_table exists. if not, skip refreshing pg metrics

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

❌ Patch coverage is 35.29412% with 11 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (stable/v4.x@d9009cf). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
src/lib/homestore_backend/hs_pg_manager.cpp	40.00%	4 Missing and 2 partials ⚠️
src/lib/homestore_backend/hs_blob_manager.cpp	28.57%	3 Missing and 2 partials ⚠️
❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@              Coverage Diff               @@
##             stable/v4.x     #409   +/-   ##
==============================================
  Coverage               ?   52.27%           
==============================================
  Files                  ?       36           
  Lines                  ?     5295           
  Branches               ?      659           
==============================================
  Hits                   ?     2768           
  Misses                 ?     2239           
  Partials               ?      288           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[xiaoxichen]

In general LGTM

1. During recovery, should we better handling destroyed PG? Currently we only partially handled the case that Index is destroyed on destroyed PG. (

   HomeObject/src/lib/homestore_backend/hs_pg_manager.cpp

   Lines 951 to 955 in 77ba5c6

   	} else {
   	RELEASE_ASSERT(hs_pg->pg_sb_->state == PGState::DESTROYED, "IndexTable should be recovered before PG");
   	hs_pg->index_table_ = nullptr;
   	LOGI("Index table not found for destroyed pg={}, index_table_uuid={}", pg_id, uuid_str);
   	}

   )

I think we should redo the HSHomeObject::pg_destroy during recovery.

2. the check index_table logic spread across the code, is it better to add an accessor into HS_PG? which checks the existence of index and logging (if the PG is destroyed) or assert.

[Besroy]

Agree, checking the PG state seems safer and cleaner. @yuwmao @koujl Do you recall the previous solution for handling crashes during destroy_pg? Perhaps we can integrate it with this fix as a unified solution

[JacksonYao287]

agree, we need to redo pg_destroy to make sure all the pg resource are cleared before pg is recreated by BR or permanently destroyed. cc @yuwmao @koujl if you want to do this, feel free to close this PR.

[yuwmao]

No, I haven't met the destroy_pg crash issue before in my memory. But also agree on redo pg_destroy.

Agree, checking the PG state seems safer and cleaner. @yuwmao @koujl Do you recall the previous solution for handling crashes during destroy_pg? Perhaps we can integrate it with this fix as a unified solution