Details
-
Bug
-
Resolution: Fixed
-
Major
-
None
-
None
-
None
Description
To ensure fast failover config polling would work correctly, I wrote a small Python script. I noted that the config poll wasn't happening as expected. Turning up the log level, I see that there is another param that needs to be set.
My script has
cluster = Cluster('couchbase://localhost?config_poll_interval=2.5')
|
.
|
27109ms [I0] {42421/1407} [TRACE] (server - L:563) <localhost:11210> (SRV=0x7ff291f26350,IX=0) Scheduling next timeout for 2499 ms. This is not an error
|
27109ms [I0] {42421/1407} [TRACE] (bootstrap - L:143) Background-polling for new configuration
|
27109ms [I0] {42421/1407} [INFO] (bootstrap - L:206) Not requesting a config refresh because of throttling parameters. Next refresh possible in 6987ms or 100 errors. See LCB_CNTL_CONFDELAY_THRESH and LCB_CNTL_CONFERRTHRESH to modify the throttling settings
|
.
|
.
|
.
|
30120ms [I0] {42421/1407} [TRACE] (server - L:563) <localhost:11210> (SRV=0x7ff291f26350,IX=0) Scheduling next timeout for 2499 ms. This is not an error
|
30120ms [I0] {42421/1407} [TRACE] (bootstrap - L:143) Background-polling for new configuration
|
30120ms [I0] {42421/1407} [INFO] (bootstrap - L:206) Not requesting a config refresh because of throttling parameters. Next refresh possible in 3976ms or 100 errors. See LCB_CNTL_CONFDELAY_THRESH and LCB_CNTL_CONFERRTHRESH to modify the throttling settings
|
.
|
.
|
.
|
33133ms [I0] {42421/1407} [TRACE] (server - L:563) <localhost:11210> (SRV=0x7ff291f26350,IX=0) Scheduling next timeout for 2499 ms. This is not an error
|
33133ms [I0] {42421/1407} [TRACE] (bootstrap - L:143) Background-polling for new configuration
|
33133ms [I0] {42421/1407} [INFO] (bootstrap - L:206) Not requesting a config refresh because of throttling parameters. Next refresh possible in 963ms or 100 errors. See LCB_CNTL_CONFDELAY_THRESH and LCB_CNTL_CONFERRTHRESH to modify the throttling settings
|
.
|
.
|
.
|
36146ms [I0] {42421/1407} [TRACE] (server - L:563) <localhost:11210> (SRV=0x7ff291f26350,IX=0) Scheduling next timeout for 2499 ms. This is not an error
|
36146ms [I0] {42421/1407} [TRACE] (bootstrap - L:143) Background-polling for new configuration
|
36146ms [I0] {42421/1407} [TRACE] (confmon - L:267) Refreshing current cluster map
|
36146ms [I0] {42421/1407} [INFO] (confmon - L:150) Not applying configuration received via CCCP. No changes detected. A.rev=24, B.rev=24
|
36146ms [I0] {42421/1407} [TRACE] (confmon - L:254) Attempting to retrieve cluster map via CCCP
|
36146ms [I0] {42421/1407} [INFO] (cccp - L:143) Re-Issuing CCCP Command on server struct 0x7ff291f26350 (localhost:11210)
|
.
|
37150ms [I0] {42421/1407} [INFO] (confmon - L:150) Not applying configuration received via CCCP. No changes detected. A.rev=24, B.rev=24
|
I'd suggest we need to coerce the other setting down if this one is adjusted. At the very least, it needs to be documented better.
Attachments
| For Gerrit Dashboard: CCBC-829 | ||||||
|---|---|---|---|---|---|---|
| # | Subject | Branch | Project | Status | CR | V |
| 82875,2 | CCBC-829: Background configuration polling should not be throttled | master | libcouchbase | Status: MERGED | +2 | +1 |