Hi Adam,
thanks for your reply. I did some heavy testing with the released 1.0.4 and
put the log in the
gist https://gist.github.com/domutom/3748f143f523a6ff95b4 named tuesday.
will try to install the Master branch now.
Here is the test sequence which I did: Please ask for more detailed testing
if you need something special.
A
20:33:52.504987-20:33:56.768576 restarting sync gateway
B
20:35:55.650127-20:36:00.788502 one change in one doc --> (no replication !)
20:37:22.568167-20:37:23.179093 delayed replication
C
20:38:40.166944-20:38:57.164195 one change in one doc --> (complete
replication)
D
20:40:57.761472-20:41:01.149174 one change in one doc --> (complete
replication)
E
20:42:14.114454-20:42:19.220205 two changes in one doc action <3sec --> (no
replication)
20:43:22.570530-20:43:22.962715 delayed replication
F
20:43:54.664340-20:44:33.412083 two changes in two docs action <3sec -->
(direct replication)
G
20:46:46.210060-20:46:49.833087 one change in one doc --> (complete
replication)
H
20:48:02.712216-20:48:05.888083 one change in one doc --> (complete
replication)
20:48:54.822455-20:48:57.915750 action not related !!
I
20:51:08.231261-20:51:13.347759 two changes in one doc action <3sec --> (no
replication)
20:52:22.573744-20:52:22.943901 replication
20:53:56.175043-20:53:58.298725 action not related !!
J
20:56:35.579062-20:56:40.480740 two changes in one doc activated on two
devices action <3sec --> (no replication)
20:57:52.576277-20:57:52.989834 replication
20:58:56.626162-20:58:58.680681 action not related !!
Am Dienstag, 7. April 2015 18:59:14 UTC+2 schrieb Adam Fraser:
>
> Interesting. I think what's actually happening here is the same issue as
> #673 - an unused sequence is being created on SG startup. However, the 60
> second timeout (waiting for #11326) doesn't start until your first update
> shows up on the the TAP feed:
>
> 02:39:08.911342 Cache: Received #11328 after 580ms
> ("59238463-B4DD-48CB-A607-D3701DBA794F" /
> "51-4a2a2fa5bce8750f1fb5d590ff24d907")
> 02:39:08.911352 Cache: Deferring #11328 (1 now waiting for #11326...#11327)
>
> You could verify this in a couple of ways:
> 1. I'd expect this to be unrelated to making multiple updates on the same
> document. Even updates on separate documents should hit this issue.
> 2. You could temporarily remove any user definitions from your config file -
> that will avoid the unused sequence.
> 3. You could test using a build of SG master branch (which has the fix for
> #673)
>
> When you originally hit this issue (the log from the first post on this
> thread), do you remember whether it was for a Sync Gateway that had been
> recently started? Since you were originally reporting this only for multiple
> updates to a single document, and not different documents, it sounds like
> there still may be more going on than just the 'unused sequence on startup'
> bug. If you're able to reproduce the original issue, I'd really like to see
> the details. One way to reproduce might be:
> - restart SG
> - send an single update (to trigger the timeout waiting for the unused
> sequence)
> - wait for the timeout (60+ seconds)
> - attempt multiple updates to a single doc, and see if you're able to
> reproduce
>
> Thanks,
> Adam
>
>
--
You received this message because you are subscribed to the Google Groups
"Couchbase Mobile" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/mobile-couchbase/8d676330-cf99-4ebd-839f-825e5da63e7b%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
