panicked at 'assertion failed: self.meta.root_page == page_id || self.page_parents.contains_key(&page_id)'
drauschenbach opened this issue · 12 comments
May 04 17:38:59 c003-n3 on-prem-agent[25208]: thread 'tokio-runtime-worker' panicked at 'assertion failed: self.meta.root_page == page_id || self.page_parents.contains_key(&page_id)', /home/REDACTED/.cargo/registry/src/github.com-1285ae84e5963aae/jammdb-0.5.0/src/bucket.rs:723:17
Can you give me any more context to reproduce this reliably?
I'll work on that. But this was all I currently have from my systemd journal where I didn't have extra stack trace info turned on.
OK it looks like this error occurs on every write attempt now, so I was able to capture the extra backtrace info.
May 04 18:12:31 c003-n3 on-prem-agent[3249]: thread 'main' panicked at 'assertion failed: self.meta.root_page == page_id || self.page_parents.contains_key(&page_id)', /home/davidr/.cargo/registry/src/github.com-1285ae84e5963aae/jammdb-0.5.0/src/bucket.rs:723:17
May 04 18:12:31 c003-n3 on-prem-agent[3249]: stack backtrace:
May 04 18:12:31 c003-n3 on-prem-agent[3249]: 0: rust_begin_unwind
May 04 18:12:31 c003-n3 on-prem-agent[3249]: at ./rustc/2fd73fabe469357a12c2c974c140f67e7cdd76d0/library/std/src/panicking.rs:493:5
May 04 18:12:31 c003-n3 on-prem-agent[3249]: 1: core::panicking::panic_fmt
May 04 18:12:31 c003-n3 on-prem-agent[3249]: at ./rustc/2fd73fabe469357a12c2c974c140f67e7cdd76d0/library/core/src/panicking.rs:92:14
May 04 18:12:31 c003-n3 on-prem-agent[3249]: 2: core::panicking::panic
May 04 18:12:31 c003-n3 on-prem-agent[3249]: at ./rustc/2fd73fabe469357a12c2c974c140f67e7cdd76d0/library/core/src/panicking.rs:50:5
May 04 18:12:31 c003-n3 on-prem-agent[3249]: 3: jammdb::bucket::BucketInner::node
May 04 18:12:31 c003-n3 on-prem-agent[3249]: at ./home/davidr/.cargo/registry/src/github.com-1285ae84e5963aae/jammdb-0.5.0/src/bucket.rs:723:17
May 04 18:12:31 c003-n3 on-prem-agent[3249]: 4: jammdb::node::Node::merge
May 04 18:12:31 c003-n3 on-prem-agent[3249]: at ./home/davidr/.cargo/registry/src/github.com-1285ae84e5963aae/jammdb-0.5.0/src/node.rs:276:39
May 04 18:12:31 c003-n3 on-prem-agent[3249]: 5: jammdb::bucket::BucketInner::rebalance
May 04 18:12:31 c003-n3 on-prem-agent[3249]: at ./home/davidr/.cargo/registry/src/github.com-1285ae84e5963aae/jammdb-0.5.0/src/bucket.rs:761:41
May 04 18:12:31 c003-n3 on-prem-agent[3249]: 6: jammdb::bucket::BucketInner::rebalance
May 04 18:12:31 c003-n3 on-prem-agent[3249]: at ./home/davidr/.cargo/registry/src/github.com-1285ae84e5963aae/jammdb-0.5.0/src/bucket.rs:748:35
May 04 18:12:31 c003-n3 on-prem-agent[3249]: 7: jammdb::bucket::Bucket::rebalance
May 04 18:12:31 c003-n3 on-prem-agent[3249]: at ./home/davidr/.cargo/registry/src/github.com-1285ae84e5963aae/jammdb-0.5.0/src/bucket.rs:354:9
May 04 18:12:31 c003-n3 on-prem-agent[3249]: 8: jammdb::transaction::TransactionInner::rebalance
May 04 18:12:31 c003-n3 on-prem-agent[3249]: at ./home/davidr/.cargo/registry/src/github.com-1285ae84e5963aae/jammdb-0.5.0/src/transaction.rs:296:9
May 04 18:12:31 c003-n3 on-prem-agent[3249]: 9: jammdb::transaction::Transaction::commit
May 04 18:12:31 c003-n3 on-prem-agent[3249]: at ./home/davidr/.cargo/registry/src/github.com-1285ae84e5963aae/jammdb-0.5.0/src/transaction.rs:166:9
May 04 18:12:31 c003-n3 on-prem-agent[3249]: 10: <on_prem_agent::configdb::jamm::ConfigDAOImpl as on_prem_agent::configdb::ConfigDAO>::store_config
...
And my calling code is:
fn store_config(&self, config: &Config) -> Result<(), Error> {
trace!("store_config({})", config.id);
let tx = self.db.tx(true)?;
let bucket = match tx.get_bucket(BUCKET_CONFIGS_BY_ID) {
Err(JammdbError::BucketMissing) => tx.create_bucket(BUCKET_CONFIGS_BY_ID)?,
Err(e) => return Err(Error::from(e)),
Ok(b) => b,
};
bucket.put(&config.id, Bytes::from(serialize_config(&config)))?;
tx.commit()?;
Ok(())Hmm, if it's happening on every write now the database file probably got into a bad state somehow. Two things that would be helpful if possible:
-
The corrupt database file itself, if you don't have anything too sensitive in there or something
-
If you know what you did up until the database got into this bad state so I could reproduce the issue from scratch.
There's nothing confidential within my DB, so you're welcome to have a copy. This is from a Raspberry Pi so the architecture was armhf, in case you need to know that to comprehend byte order.
I'll need some time but I'll try to dig through the log history to find the initial failure event.
As for what I was doing when the database became corrupt, a code review reveals that the code shown above is the only thing going on on a regular basis (besides read-only ops which I'm assuming are irrelevant).
I am also having this problem in my application.
Unfortunately I am unable to share my database at the point of the issue, but it occurs consistently with a particular sequence of writes.
My code looks like this:
fn put(&mut self, key: u64, value: &[u8]) -> std::result::Result<(), Error> {
let tx_result = (|| {
let tx = self.db.tx(true)?;
let bucket = tx.get_or_create_bucket("cache")?;
let r = bucket.put(key.to_be_bytes(), value).map(|_| ());
tx.commit()?;
r
})();
// ... other code
tx_result
}Hey @brandonhamilton, I know there is a race condition somewhere in here but I'm not able to reproduce it consistently so I'm having a hard time finding it. Do you have steps that consistently reproduce the error?
Same issue for me too
thread 'actix-rt|system:0|arbiter:0' panicked at 'assertion failed: self.meta.root_page == page_id || self.page_parents.contains_key(&page_id)', /home/admins/.cargo/registry/src/github.com-1ecc6299db9ec823/jammdb-0.5.0/src/bucket.rs:723:17
cargo 1.61.0-nightly (109bfbd 2022-03-17)
release: 1.61.0-nightly
host: x86_64-unknown-linux-gnu
os: Ubuntu 20.04 (focal) [64-bit]
---------Code
match jdb.tx(true) {
Err(err) => Ok(HttpResponse::BadGateway().body(err.to_string())),
Ok(tx) => {
match tx.get_bucket("jsmaster") {
Err(err) => Ok(HttpResponse::BadGateway().body(err.to_string())),
Ok(pages) => {
match pages.put(id.clone().as_bytes(),json::stringify(parsed)){
Err(err) => Ok(HttpResponse::BadGateway().body(err.to_string())),
Ok(_) => {
match tx.commit() {
Err(err) => Ok(HttpResponse::BadGateway().body(err.to_string())),
Ok(_) => Ok(HttpResponse::Ok().body(id))
}
}
}
}
}
}
}
I'm also reproducing this with 100% reproducibility and a fresh file each time.
@abc-mikey Can you share your code and your system details?
I believe 0.8.0 should resolve this issue, feel free to open another one if you run into other problems!