gcache and recv-queue

[Yos Tj]

Hi,

Would you please tell me in detail about the difference between
gcache and recv-queue, so that I can understand these parameters?

gcache.size (128MB as default)
gcs.fc_limit (16)
gcs.recv_q_hard_limit (LLONG_MAX)
gcs.recv_q_soft_limit (0.25)

Regards,

[Philip Stoev]

Hello,

gcache.size constrains the total size of replication events that a node
keeps on disk in case it needs to provide those events to other nodes that
have temporarily dropped from the cluster, the so called IST, Incremental
State Transfer.

gcs.recv_q_hard_limit constrains the total size of incoming replication
events a node will keep while it is undergoing a full database transfer, the
so called SST , State Snapshot Transfer. While a node is fetching the entire
database from another node, new transactions will keep arriving and the node
needs to store them until the SST is complete.

gcs.fc_limit defines the number of transactions a node is allowed to fall
behind before the flow control mechanism kicks in and the entire cluster is
throttled until the node can catch up.

--
You received this message because you are subscribed to the Google Groups
"codership" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to codership-tea...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[James Wang]

Any tool to read the galera.cache file please?

Thanks in advance

[Yos Tj]

Hello Philip,

Now I understand they are quite different.

gcache preserves outgoing transactions,
in case other nodes need IST.

recv-queue preserves incoming transactions,
so that this node can apply them after finishing SST.

Thanks a lot!!

Regards,