# TiKV Crashes with SIGSEGV on ARM Weak-Memory Platforms

Source note: adapted from the Feishu `Storage oncall handbook` <https://pingcap.feishu.cn/wiki/RNzIwCexFiA1uXkRCvzcsIocnVb>.

- Phenomenon:
  - One or more TiKV nodes crash with `SIGSEGV`.
  - In severe cases, many TiKV instances can crash within a short time window.
- Trigger / Preconditions:
  - The environment uses ARM weak-memory behavior.
  - The failure is associated with an upstream RocksDB race in `MemTable::GetBloomFilter()`.
- Key logs:
  - The exact backtrace can vary. Treat `SIGSEGV` together with RocksDB stack frames as the primary signal.
- Key metrics:
  - This is primarily a crash-class incident rather than a metrics-first incident.
  - Node or cluster availability degrades through restart loops or concurrent process exits.
- Root cause:
  - An upstream RocksDB issue can trigger a data race in `MemTable::GetBloomFilter()`.
  - The problem is especially relevant on ARM weak-memory platforms and is difficult to reproduce consistently.
- Workaround:
  - Collect core dumps when possible.
  - Treat the incident as version- and platform-sensitive; prepare upgrade or rollback decisions based on the affected release line.
- Fixed version:
  - Follow the upstream fix status for the RocksDB-side issue before relying on version-only mitigation.
- Related oncall:
  - v8.5.3 Segment fault in RocksDB write operation `ONCALL-8053`
  - Manus v8.5.1 TiKV segfault `GTOC-7804` `APID-11531`
  - Blot v8.5.1 57/63 TiKV instances crashed in two minutes `EMID-10050` `GTOC-7770`
  - Airbnb v8.5.1 TiKV segmentation violation (`SIGSEGV`) crash `GTOC-7785` `NAID-11514`