tiflow: Compare with v6.6.0, v7.0.0 has 140%-180% performance regression with errors on sysbench oltp_write_only workload
What did you do?
- deploy 2 TiDB and cdc
- run sysbench oltp_write_only
What did you expect to see?
No performance regression, the cdc_changefeed_checkpoint_lag is similar.
What did you see instead?
Versions of the cluster
Upstream TiDB cluster version (execute SELECT tidb_version();
in a MySQL client):
(paste TiDB cluster version here)
v7.0.0
Upstream TiKV version (execute tikv-server --version
):
(paste TiKV version here)
v7.0.0
TiCDC version (execute cdc version
):
(paste TiCDC version here)
v7.0.0
About this issue
- Original URL
- State: closed
- Created a year ago
- Comments: 15 (11 by maintainers)
Commits related to this issue
- sink(ticdc): add max allowed packet parmameters in sink dsn (#8627) close pingcap/tiflow#8622 — committed to pingcap/tiflow by amyangfei a year ago
- sink(ticdc): add max allowed packet parmameters in sink dsn (#8627) (#8632) close pingcap/tiflow#8622 — committed to pingcap/tiflow by ti-chi-bot a year ago
- sink(ticdc): turn off multi statement sql by default (#8652) close pingcap/tiflow#8622 — committed to pingcap/tiflow by amyangfei a year ago
- sink(ticdc): add max allowed packet parmameters in sink dsn (#8627) close pingcap/tiflow#8622 — committed to 3AceShowHand/tiflow by amyangfei a year ago
We met many packet errors in dmesg log
The message
enp189s0f0: L3/L4 error pkt
indmesg -T
indicates that there was a Layer 3/Layer 4 (L3/L4) error packet encountered on network interfaceenp189s0f0
.Layer 3 refers to the network layer in the OSI model, which is responsible for routing and forwarding of data packets between different networks. Layer 4 refers to the transport layer in the OSI model, which provides end-to-end communication services and ensures the reliable delivery of data between hosts.
The error message suggests that there was an issue with the packet’s routing or delivery at the network or transport layer. This could be due to various reasons such as incorrect configuration, network congestion, faulty hardware, or software issues.
We should investigate the cause of the error further to determine if there is a problem with the network infrastructure or the configuration.