LiteFS Panic when WAL is large #445
Description
Hi, I recently tried vacuuming my database which produced a large WAL file (~1GB) to be checkpointed. LiteFS has been repeatedly outputting this error after this about a panic and the WAL does not get checkpointed.
level=INFO msg="fuse: panic in handler for Flush [ID=0x57ecca Node=0x4 Uid=0 Gid=0 Pid=3291] 0x3 fl=0x0 owner=bcd568aeef4157ee: runtime error: index out of range [970] with length 968\ngoroutine 2887718 [running]:\nbazil.org/fuse/fs.(*Server).serve.func2()\n\t/go/pkg/mod/bazil.org/[email protected]/fs/serve.go:1004 +0x119\npanic({0xaafd00?, 0xc000400090?})\n\t/usr/local/go/src/runtime/panic.go:914 +0x21f\ngithub.com/superfly/litefs.(*DB).checksum(0xc000188dc0, 0x3c752, 0x3cc61?)\n\t/src/litefs/db.go:3233 +0x42c\ngithub.com/superfly/litefs.(*DB).CommitWAL(0xc000188dc0, {0xbec848, 0xc0001d6c30})\n\t/src/litefs/db.go:1687 +0x1bf7\ngithub.com/superfly/litefs.(*DB).UnlockSHM(0xc000188dc0, {0xbec848, 0xc0001d6c30}, 0xc0000388b0?)\n\t/src/litefs/db.go:1873 +0x87\ngithub.com/superfly/litefs/fuse.(*SHMHandle).Flush(0xc000210000?, {0xbec848?, 0xc0001d6c30?}, 0xc000747d40?)\n\t/src/litefs/fuse/shm_node.go:144 +0x28\nbazil.org/fuse/fs.(*Server).handleRequest(0xc000210000, {0xbec848, 0xc0001d6c30}, {0xbe7660?, 0xc0000388a0}, 0xc0001c22c0, {0xbebc00?, 0xc0001d6be0?}, 0xc0007bff08)\n\t/go/pkg/mod/bazil.org/[email protected]/fs/serve.go:1461 +0x190b\nbazil.org/fuse/fs.(*Server).serve(0xc000210000, {0xbebc00, 0xc0001d6be0})\n\t/go/pkg/mod/bazil.org/[email protected]/fs/serve.go:1025 +0x5b1\nbazil.org/fuse/fs.(*Server).Serve.func1()\n\t/go/pkg/mod/bazil.org/[email protected]/fs/serve.go:522 +0x5a\ncreated by bazil.org/fuse/fs.(*Server).Serve in goroutine 76\n\t/go/pkg/mod/bazil.org/[email protected]/fs/serve.go:520 +0x373"