Known high-priority issues from LizardFS

[uristdwarf]

These are know high-priority issues the SaunaFS team has analyzed from LizardFS that still potentially exist.

• EC goals are not safe in certain circumstances: when same IP chunkserver stores more than 1 part, if the disk on that chunkserver fails then data is lost. Related:
  XOR and EC goals are not safe, even when just single disk fails lizardfs/lizardfs#910

• In uraft, a shadow with older metadata may become master first when starting up. Related:
  lizardfs-uraft failed to start lizardfs/lizardfs#824
  New shadow server join into cluster，but load medata failed； lizardfs/lizardfs#682

• Chunks may be lost after snapshot is deleted. Related:
  Chunks Lost When Snapshots Are Removed lizardfs/lizardfs#799

• Potentially slow write speeds. Related:
  Very slow write speeds lizardfs/lizardfs#710,
  Single Threaded Write Performance Question/Issue lizardfs/lizardfs#703

If you find an issue that was missed and still exists on SaunaFS, please open a new issue instead, thank you!

[uristdwarf]

Chunks may be lost after snapshot is deleted. Related:

This part I wasn't able to reproduce in SaunaFS, so I will consider closed until someone is able to reproduce it again.