Hi, I'm expiriencing an issue with HelpServer. Sometimes after some period of idle time (nobody uses HS) in response of a next request HS starts rebuilding its index. But no one of its resources (chm-files) hasn't changed. And there aren't recycling in IIS (IIS7 on WinSrv2008 x64). Why does it do this? After a page has rendered w3wp process with HS's AppPool continues consuming CPU. Ok, it may be some async reindexing, but... Working set at that time is ~250Mb for hosting one 30Mb chm-file. Is it normal? After more time goes and next request make HS rebuilds index, working set is ~500Mb (!) for the same single 30Mb CHM-file (it physical memory! not just virtual VM). I guess it leaks somewhere... I'm sure 500Mb for hosting a 30Mb's chm-file is too much. using IIS7 on WinServer 2008 x64, AppPool is in x64 integrated mode This thread was imported from our support forum. The original discussion may contain more detailed answer. Original topic by EvilShrike. |
> in response of a next request HS starts rebuilding its index. Actually it reads the index ("rebuild" = "read"). This happens on web application startup, or after recycling. > Ok, it may be some async reindexing, but... Exactly. It builds full-text indexes in background after showing its "main window" (TOC-Index-Search page). > I'm sure 500Mb for hosting a 30Mb's chm-file is too much. Yes... Even taking into account 30Mb ~= 300-400Mb of uncompressed HTML. We'll study the issue. We've just shared ~ 100Mb of compressed help files on http://help-server.com/ , which is hosted @ ReliableSite.net First 15 minutes were hard for it: it seems application was several times recycled, but our index update process is ready for this. But after this period it operates smoothly. Ok, I don't mean we won't study the issue. I just wanted to show it actually doesn't lead to serious problems even with very big files. |