Getting Bad fetch errors on VCS

[chandra]

I got a issue with the badfetch errors on VCS servers. Every time i get the problem i need to clear the cache of the vcs and restart it again.The problem is that we have total of 9 vcs servers, but we get the problem on different vcs servers everytime. Is there a way that we can clear the cache in an auotmated way? and any expected solutions for this bad fetch.Help is greatly appreciated.

Thanks,
Chandra

[Genesyde]

Here's how I clear our cache.  You can create this in a batch file then setup an scheduled task at whatever interval you need.

del /Q C:\GVP\cn\web\cache\localhost\VPM\Database\AppUrl\DTMF_VOX\*.*
del /Q C:\GVP\cn\web\cache\localhost\VPM\Database\AppUrl\Speech_VOX\*.*
del /Q C:\GVP\cn\web\cache\localhost\VxmlRoot\VoxFiles\en-us\*.*

[chandra]

Genesyde,

Thanks for the reply. Does we need to restart the watch dog? We are using windows environment. Does it effect the calls during this process?

Thanks,
Chandra

[chandra]

Thanks for the quick posts.

Still tweaking  with this. Our configuration is IVR<->F5(load balancer)<-> WAS(websphere servers). My findings till now are during badfetch errors some GVp is getting 1/8 th of the total file i.e partial pages being delivered. Once we restart the watch dog, the cache get cleared and works fine for 2 days. Then again the problem comes around everycouple of days. Temp files are being loaded in GVP\CN\temp folder at this error. After taking a network trace our network folks say that nothing is related to network side.
I am trying hard to resolve this issue but this seems to be night mare....Any ideas please advise me in trouble shooting.....

Thanks,
chandu123

[Mine]

No good way.
Can you change parameter "Cache file function is off"?

[chandra]

Hi All,

I finally found the wht is causing these badfetch messages. The problem is with the static page which should not be cached are being cached. Because of this unusal caching mechanism cached files are hitting the max limit ( According to genesys limit is either 16478(like that)  or 4 GB) and validation is taking place, which results in a partial page . We modified the caching behaviour which solved this problem.

Thanks for the support....

--Chandra