Boein said:
For the moment I only seem to have the problem with our intranet website.
This website is running on an apache server using typo3. I haven't found an
external website yet, but on our internal I can very easily reproduce the
problem. I also did some monitoring with filemon and regmon.
I wouldn't expect to see anything much of interest from RegMon
for this problem (unless you were really lucky and were able to identify
where the new assumed 1 minute rendering timeout originates <w>).
FileMon filtering on iexplore*temp would have shown you the writes
being done in the TIF as the images were being downloaded from the
web. Alternatively, if the images were cached and the the server had
sent back an acknowledgement that it was Ok to use the cached copy
you would have seen the reads for those files. More interesting but
harder to recognize would be the case where the server Ok'd the use
of the cached copy but then there was no read in the TIF. That was
one example of a Red-X that I once saw (a long time ago) in a late
version of IE5 and an early version of IE6.
Most Red-X these days have much simpler explanations which almost
always involve external factors which you can detect by an HTTP trace
using a tool such as FiddlerTool.
Also the guys from MS sent me and logging tool.
NetMon perhaps? If you have the XP Support tools installed
you can get almost the same type of trace using netcap.
Netcap creates .cap files which can then be analysed and formatted
by a program such as Ethereal. Ethereal can do its own capture too
but I find it harder to use for that and less transparent than netcap.
After analizing it they came up with this:
"There is one thing a little bit strange in there: after you hit the back,
there is no “GET†request for the picture, but the server sends an ok for
it."
Sounds like a record was lost. <eg>
Was there a CPU spike at the same time?
I don't know what they mean by that but maybe this helps for you guys?
Also I did a test with Firefox, installed it on the same server... problem
solved when going to the website with mozilla... red cross when going to the
website with IE7. I really want to use IE but if I don't find a solution
very vast, firefox I will have to install Firefox.
I reported Red-X during the beta for images which IE and the
IE Developer Toolbar refused to render. Perhaps you are seeing
one of those? In those cases, OE could capture them and then
view them with the Windows Picture and Fax Viewer.
E.g. try using right-click, E-mail Picture... on the Red-X
and then if OE downloads or otherwise puts the image into an attachment
doubleclick on the attachment and see if it opens properly.
Oh. I've just re-read the whole thread. Apparently this Red-X is intermittent
for (ostensibly) the same file? In that case using the FileMon trace make sure
that the size of the file being read is the same each time. E.g. I think
that the reason for the Red-X that I saw (the ones that both IE7 and the
IE Developer Toolbar didn't like but OE could use) were caused by
extraneous information which had been added to the bottom of the file
(e.g. because the file had been created by a trial version of some
shareware which tacked on some words to that effect at the end of
the image file.)
Another possible explanation would depend on whether your host
has multiple servers (e.g. in a cluster of clones). Sometimes such
clones get out of synch, so that some have an image, some have
an older version of it, or some don't have it all. It would be the latter
case which could result in an intermittent Red-X but then that should
be easily detectable in an HTTP trace by an 404 response.
HTH
Robert
---
