Quantcast
Channel: StackExchange Replication Questions
Viewing all articles
Browse latest Browse all 17268

seqno in grastate.dat (Galera) - does seqno increment for every transaction? Are we 300 million transactions behind?

$
0
0

After storage outage those are the sequence numbers of our cluster:

galera-1: 515990451
galera-2: 214013741
galera-3: -1

Is this correct? On galera-1 is the most recent data? According to this question the most recent data state is on galera-3 with seqno -1?

When looking at galera-1 and galera-2: did galera-2 missed 301976710 transactions? It means not in sync? What are reasons that this happens? This is very huge inconsistency.

the original file

# cat /var/lib/mysql/data/grastate.dat
# GALERA saved state
version: 2.1
uuid:    b7677bae-d703-11e6-9df6-ee19cbf30410
seqno:   -1
cert_index:

our versions / RHEL7 with Galera from OpenStack channel. Sorry for the legacy versions, we are in progress to upgrade that to 10.x

> show status like 'wsrep_provider_version';
+------------------------+------------+
| Variable_name          | Value      |
+------------------------+------------+
| wsrep_provider_version | 3.5(rXXXX) |
+------------------------+------------+
1 row in set (0.00 sec)

> select VERSION();
+----------------------+
| VERSION()            |
+----------------------+
| 5.5.41-MariaDB-wsrep |
+----------------------+
1 row in set (0.00 sec)

This posting at Percona forums have similar questions like me.

I'm trying to understand the entries in grastate.dat file, particularly seqno value. If I understand correctly, the 'seqno' value increases as node gets written to. If all nodes are shut down I can use 'seqno' to figure out which node was last written to - it is the one which has the highest 'seqno' value.


Viewing all articles
Browse latest Browse all 17268

Trending Articles