Known issue with Badger? sync.runtime_Semacquire panic

Users
1

I’m running a sustained load test using badger (no dgraph just yet) to test the underlying scalability. I’m writing ~50KEPS on an i3.4XL in AWS and after around 24 hours with a 12 hour TTL the program panic’d with the below paste. Just seeing if this is a known issue before making a more targeting test case to post.

opts := badger.DefaultOptions
opts.Dir = config.DBDirectory
opts.ValueDir = config.DBDirectory
opts.TableLoadingMode = options.MemoryMap
opts.MaxTableSize = 2 * (1024 * 1024 * 1024) // 2GB

And a go routine running every 30 mins on a ticker that calls: (err handling removed for clarity)
g.db.PurgeOlderVersions()
g.db.RunValueLogGC(0.5)

goroutine 2787863423 [semacquire]:
sync.runtime_Semacquire(0xc7bfdec26c)
        /usr/local/go/src/runtime/sema.go:56 +0x39
sync.(*WaitGroup).Wait(0xc7bfdec260)
        /usr/local/go/src/sync/waitgroup.go:129 +0x72
github.com/dgraph-io/badger.(*request).Wait(0xc7bfdec230, 0x460261, 0xc4205ae1e0)
        /mnt/mrnull/golang/src/github.com/dgraph-io/badger/value.go:712 +0x35
github.com/dgraph-io/badger.(*DB).batchSetAsync.func1(0xc7bfdec230, 0xc9205b39a0)
        /mnt/mrnull/golang/src/github.com/dgraph-io/badger/db.go:727 +0x2b
created by github.com/dgraph-io/badger.(*DB).batchSetAsync
        /mnt/mrnull/golang/src/github.com/dgraph-io/badger/db.go:726 +0x9c

goroutine 2787884118 [running]:
        goroutine running on other thread; stack unavailable
2

No, this isn’t a known issue. It would be helpful if you can share the entire stack trace and a program which reproduces this.

3

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.