Closed dpage closed 2 years ago
Comment migrated from Redmine: https://redmine.postgresql.org/issues/2556#note-1 Originally created by Ashesh Vashi at 2017-07-13 14:16:36 UTC.
Removing Neel as assignee, as he is already having some cases assigned to him.
Redmine ticket header update:
Name | Old Value | New Value |
---|---|---|
Assigned To changed | Neel Patel |
Comment migrated from Redmine: https://redmine.postgresql.org/issues/2556#note-2 Originally created by Khushboo Vashi at 2017-07-14 05:51:25 UTC.
Redmine ticket header update:
Name | Old Value | New Value |
---|---|---|
Assigned To changed | Khushboo Vashi |
Comment migrated from Redmine: https://redmine.postgresql.org/issues/2556#note-3 Originally created by Khushboo Vashi at 2017-07-17 08:24:30 UTC.
Redmine ticket header update:
Name | Old Value | New Value |
---|---|---|
Status changed | New | In Progress |
Comment migrated from Redmine: https://redmine.postgresql.org/issues/2556#note-4 Originally created by Khushboo Vashi at 2017-07-17 09:50:11 UTC.
Hi Dave,
Ensure we retry at least once once the timeout has elapsed. This means after 30 seconds, one more try to ping the server, is this correct? If yes then why we need to retry as we have already try for 30 seconds in a loop?
Make the timeout a config option (not in the UI, just the registry/config file). My understanding is, at first time update the registry with the timeout parameter and then use that every time. But if no UI, then do we need to intimate the user on failure that the connection timeout can be changed through the registry?
Thanks, Khushboo
Redmine ticket header update:
Name | Old Value | New Value |
---|---|---|
Assigned To changed | Khushboo Vashi | Dave Page |
Comment migrated from Redmine: https://redmine.postgresql.org/issues/2556#note-5 Originally created by Dave Page at 2017-07-17 10:08:26 UTC.
Khushboo Vashi wrote:
Hi Dave,
- Ensure we retry at least once once the timeout has elapsed. This means after 30 seconds, one more try to ping the server, is this correct? If yes then why we need to retry as we have already try for 30 seconds in a loop?
We retry in a loop so that we can start as quickly as possible. The last attempt (outside the loop) is to ensure that if there was a long network timeout whilst we were looping that resulted in only the initial attempt being made, we can still recover from that.
- Make the timeout a config option (not in the UI, just the registry/config file). My understanding is, at first time update the registry with the timeout parameter and then use that every time.
No, just attempt to read the registry, but have a default value. For example:
int timeout = settings.value("ConnectionTimeout", 30).toInt();
But if no UI, then do we need to intimate the user on failure that the connection timeout can be changed through the registry?
No - but we probably should add a doc page containing all the configuration settings and what they are. We have 3 or 4 now.
Redmine ticket header update:
Name | Old Value | New Value |
---|---|---|
Assigned To changed | Dave Page | Khushboo Vashi |
Comment migrated from Redmine: https://redmine.postgresql.org/issues/2556#note-6 Originally created by Khushboo Vashi at 2017-07-17 12:54:49 UTC.
The patch submitted.
Comment migrated from Redmine: https://redmine.postgresql.org/issues/2556#note-7 Originally created by Khushboo Vashi at 2017-07-17 14:18:03 UTC.
Applied in changeset commit:c5461827894bc99a9d4909c078397a30a53dfee5.
Redmine ticket header update:
Name | Old Value | New Value |
---|---|---|
Status changed | In Progress | In Testing |
Done Ratio changed | 0 | 90 |
Comment migrated from Redmine: https://redmine.postgresql.org/issues/2556#note-8 Originally created by Fahar Abbas at 2017-09-08 11:10:05 UTC.
This is resolved in Desktop Runtime installer on MAC and Windows.
Redmine ticket header update:
Name | Old Value | New Value |
---|---|---|
Status changed | In Testing | Resolved |
Comment migrated from Redmine: https://redmine.postgresql.org/issues/2556#note-9 Originally created by Fahar Abbas at 2018-01-09 05:52:13 UTC.
Redmine ticket header update:
Name | Old Value | New Value |
---|---|---|
Done Ratio changed | 90 | 100 |
Issue closed on Redmine.
Issue migrated from Redmine: https://redmine.postgresql.org/issues/2556 Originally created by Dave Page at 2017-07-12 11:37:31 UTC.
We should change the runtimes connection retry loop timeout so that it's time based rather than count based. This is because some connection failures can be near instant (connection actively refused) whilst others can take time. Per a brief discussion with Ashesh:
Change the timeout to be time based, e.g. 30 seconds.
Ensure we retry at least once once the timeout has elapsed.
Make the timeout a config option (not in the UI, just the registry/config file).