Re: Extending the “No Instance Available” timeout
Re: Extending the “No Instance Available” timeout
- Subject: Re: Extending the “No Instance Available” timeout
- From: OC <email@hidden>
- Date: Wed, 06 Apr 2016 18:02:58 +0200
Aha, I see — I thought Eclipse uses Direct Connect, just as Xcode used to (and mine still does).
In that case, beside the adaptor log, I would probably
(a) try Direct Connect; no timeouts in this mode;
(b) add extra logs to show when the R/R loop begins, how it runs, and when it ends, to see what takes that long (and how long precisely).
All the best,
OC
On 6. 4. 2016, at 17:55, Lon Varscsak <email@hidden> wrote:
> Just to help Ben while he’s sleeping… This is while he’s running the application from within Eclipse (still through Apache) while he’s testing/debugging.
>
> -Lon
>
> On Wed, Apr 6, 2016 at 8:37 AM, OC <email@hidden> wrote:
> Benjamin,
>
> On 6. 4. 2016, at 10:19, Benjamin Chew <email@hidden> wrote:
>
> > I did change Receive Timeout to 999,999 under “Site”, and restarted wotaskd and monitor, but that didn’t help.
>
> The only thing which should need to be restarted is your application.
>
> > I also tried creating an app in monitor with the same name as the one running in Eclipse, and changing the Receive Timeout, but that didn’t work either.
>
> I am afraid I do not quite get your setup; what has Eclipse to do with that?
>
> Anyway, it is really weird if setting a super-high receive timeout for an application does not affect that application's “No Instance Available“ report; that does not make any sense to me. If the application locks, you still might not get the desired page, but (a) that is extremely improbable if it does run all right in a different setup, (b) at the very least, you should be waiting for the “No Instance Available“ report much, much longer.
>
> Sorry for an extremely dumb question, but is there perhaps any possibility you are changing the timeouts of another application, or the same application but on a different server, or something like that?
>
> Is there anything of interest in the adaptor log?
>
> All the best,
> OC
>
>
> > On Wed, Apr 6, 2016 at 2:59 PM, Chuck Hill <email@hidden> wrote:
> > Receive Timeout is set in JavaMonitor.
> >
> > From: Benjamin Chew <email@hidden>
> > Date: Tuesday, April 5, 2016 at 11:52 PM
> > To: Chuck Hill <email@hidden>
> > Cc: OC <email@hidden>, WebObjects-Dev Mailing List <email@hidden>
> > Subject: Re: Extending the “No Instance Available” timeout
> >
> > Thanks guys, I appreciate all the other suggestions, but I’m quite positive it is the ping time that is killing me.
> >
> > I have tried multiple apps, and all the apps have the same problem. They worked fine when I was in the US, but I encountered this problem once I got to Singapore. I’ve verified this by pinging servers in the US, and I’m getting ~300ms ping times, with some jitter, which does not help.
> >
> > OC and Chuck: could you tell me how to adjust my Receive Timeout?
> >
> > Thanks,
> > Ben
> >
> > On Wed, Apr 6, 2016 at 2:26 AM, Chuck Hill <email@hidden> wrote:
> > I assume that you are running the app locally through Apache as that message is from wotaskd. As OC pointed out, the Receive Timeout is what you need to adjust up and up and up.
> >
> > It sounds like latency is what is killing you, I don’t recall how chatty JDBC is but it is probably along the lines of ODBC which is quite chatty indeed. Latency kills its performance. Another possibility is to run a local copy of the DB.
> >
> > Chuck
> >
> >
> >
> >
> > On 2016-04-05, 7:38 AM, "webobjects-dev-bounces+chill=email@hidden on behalf of OC" <webobjects-dev-bounces+chill=email@hidden on behalf of email@hidden> wrote:
> >
> > >Benjamin,
> > >
> > >On 5. 4. 2016, at 11:02, Benjamin Chew <email@hidden> wrote:
> > >
> > >> I’m in Singapore working off a VPN connection to the States, and while waiting for some database-intensive components to display, I keep getting “No Instance Available” because it’s taking so long to complete all the queries (ping times ~ 200ms).
> > >
> > >As others have pointed out, ping times could hardly affect this.
> > >
> > >> I’ve tried going to WOMonitor on my local machine (localhost:56789) and modified the Send, Receive and Connect timeouts, but that didn’t seem to help.
> > >
> > >Far as I can say with my very limited knowledge,
> > >
> > >(a) “No Instance Available” is most time (if not always) caused by the receive timeout at the server side;
> > >(b) and thus, increasing it enough should help.
> > >
> > >> Does anyone have any ideas?
> > >
> > >First thing, I would try some ludicrously high receive timeout. For us, it always helped (in the sense that the rendered page did always return, presumed the user had the patience to wait long enough, especially when by a mistake I had computed some results in O(2^N) :))
> > >
> > >It might also help to check the adaptor log -- touch /tmp/logWebObjects as root, and the log should appear in /tmp/WebObjectsLog.
> > >
> > >The ultimate solution, of course, would be background processing and/or paging, as others already recommended; but first you need to find the particular cause of the long processing, which might be sometimes a bit hairy.
> > >
> > >All the best and good luck,
> > >OC
> > >
> > >
> > > _______________________________________________
> > >Do not post admin requests to the list. They will be ignored.
> > >Webobjects-dev mailing list (email@hidden)
> > >Help/Unsubscribe/Update your Subscription:
> > >
> > >This email sent to email@hidden
> >
> >
>
>
> _______________________________________________
> Do not post admin requests to the list. They will be ignored.
> Webobjects-dev mailing list (email@hidden)
> Help/Unsubscribe/Update your Subscription:
>
> This email sent to email@hidden
>
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden