hcq/MaxScale
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

MXS-2144: Treat server shutdown as a network error

Browse Source 
If the server where a query is being executed is shutting down,
readwritesplit should treat it as an error to make retrying of the query
possible.

By treating server shutdowns as network errors, the same code path that is
used for actual network errors can be taken. This removes the need for any
extra retrying logic for this particular case.
This commit is contained in:
Markus Mäkelä
2018-11-13 17:20:50 +02:00
parent 37d38593d5
commit cab8a4bde8
3 changed files with 38 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
server/modules/routing/readwritesplit/rwsplitsession.cc
View File

@ -555,6 +555,12 @@ void RWSplitSession::manage_transactions(SRWBackend& backend, GWBUF* writebuf)
}
}
static bool server_is_shutting_down(GWBUF* writebuf)
{
uint64_t err = mxs_mysql_get_mysql_errno(writebuf);
return err == ER_SERVER_SHUTDOWN || err == ER_NORMAL_SHUTDOWN || err == ER_SHUTDOWN_COMPLETE;
}
void RWSplitSession::clientReply(GWBUF* writebuf, DCB* backend_dcb)
{
DCB* client_dcb = backend_dcb->session->client_dcb;
@ -578,6 +584,14 @@ void RWSplitSession::clientReply(GWBUF* writebuf, DCB* backend_dcb)
}
return;
}
else if (backend->get_reply_state() == REPLY_STATE_START && server_is_shutting_down(writebuf))
{
// The server is shutting down, act as if the network connection failed. This allows
// the query to be retried on another server without the client noticing it.
poll_fake_hangup_event(backend_dcb);
gwbuf_free(writebuf);
return;
}
if ((writebuf = handle_causal_read_reply(writebuf, backend)) == NULL)
{