Hi,
After having some performace issues with 7.2.1, we set up a network sniff to try and see what was happening.
One of the interesting things we saw was that their seems to be continual handshaking between the client and the web server while running a report.
We have the polling interval set to 10 seconds, but are getting handshakes at 8 times per second.
Also the chunk size is a maximum of 1.5K rather than more optimal 4k chunks - i.e. the header is split in to three - compounding the problem.
I have asked gold support, but they are not very forth coming (they have the sniff output and are still looking at it a month down the line).
Does anyone know if it is possible to change the chunk size to a more optimal 4K.
Note: this makes no difference if web is in Get or Put mode (i.e. all on a single header - bigger and thus more pieces, or on two headers and chopoped half as much).
I don't think this has anything to do with the performace problems, per say, though it may explain why the web servers (load balanced dual 1.4Mhz 4GB machines!) are maxing out at 300 or so active users - just too much traffic!
(I have also launched a thread to do with the client performance issue, so light on it here).
Cheers,
Wolf
After having some performace issues with 7.2.1, we set up a network sniff to try and see what was happening.
One of the interesting things we saw was that their seems to be continual handshaking between the client and the web server while running a report.
We have the polling interval set to 10 seconds, but are getting handshakes at 8 times per second.
Also the chunk size is a maximum of 1.5K rather than more optimal 4k chunks - i.e. the header is split in to three - compounding the problem.
I have asked gold support, but they are not very forth coming (they have the sniff output and are still looking at it a month down the line).
Does anyone know if it is possible to change the chunk size to a more optimal 4K.
Note: this makes no difference if web is in Get or Put mode (i.e. all on a single header - bigger and thus more pieces, or on two headers and chopoped half as much).
I don't think this has anything to do with the performace problems, per say, though it may explain why the web servers (load balanced dual 1.4Mhz 4GB machines!) are maxing out at 300 or so active users - just too much traffic!
(I have also launched a thread to do with the client performance issue, so light on it here).
Cheers,
Wolf