"disk read error" during Publish to Web

[justinhill]

Guys,

I've started getting an error message

"Disk I/O error during read."

during an upload. The thing is, I don't think there really is a hard problem reading the drive - partly because I've tested it every way I can, including copying the site1 contents back and forth, and partly because Actinic is running on Windows Vista inside Parallels on a Mac so the 'drive' is the same one the Mac O/S is using and it gives everything a clean bill of health.

So. I'm thinking its some kind of permissions thing or maybe a corrupted image file but I don't know how to get Actinic to tell me which file(s) are generating the error. The message comes up while the check+send is still going on behind it.

Is there a way to get Actinic to output a log of file operations so I can pin down what's happening?

Thanks
Justin

[NormanRouxel]

No log that I know of.

Simplest way to track down a needle in haystack is with a binary chop.

Temporarily delete half your Sections and test.

If all OK, undelete, temp delete the other half and test again.

Once you know which block is faulty, half that block and repeat.

Eventually you end up with only a single item.

If you have 1000 sections you'll only need about 10 iterationss as you half the problem each time.

[justinhill]

No log. Hmmm. Binary chop is a plan but I was hoping to avoid test uploads since they'll affect the live site.

Justin

[completerookie]

with any disk i/o error, whether virtual or real , I always do the following:

• copy the enire actinic folder (everything) to a new place (ie brand new folder probably called "copy of actinic")
• rename the top of the original actinic to "actinic - corrupt"
• create a new folder called actinic
• copy the saved data back to the new folder

doing it this way ensures that the "faulty area" is still recorded as faulty (and will never will get used again) and all the data is put back into a brand new area of the hard disk hoping that the new area will be fault-free

the problem is that once the data is generating a 'faulty io' error, its difficult to guarantee that the data will always be available to you,
naturally, if you have any shared (mapped) areas, you will need to re-share them

[Darren B]

No log. Hmmm. Binary chop is a plan but I was hoping to avoid test uploads since they'll affect the live site.

Justin

switch to test mode to perform this - your trying to see whats wrong not if a live site looks good and works

[NormanRouxel]

If you think this is O/S rather than hardware related or a corrupt bit of Actinic data, then you might want to check that's nothing is indexing files or automatically backing them up on the Vista VM. A file being opened for indexing / backup when Actinic is just about to write to it would cause an error.

Also you could try copying the site to another computer and see what happens when you do an upload from there (using Test mode may be a good idea).

[justinhill]

Thanks everyone for your input - Test Mode does not produce the same errors.

In Test Mode, if I choose 'Refresh', it goes right through to the last file ('discount.fil') and then hangs until I press cancel - at which point it produces the disk error message.

In Live Mode, it does everything described above AND produces a message near the beginning of the upload process AND about half way through the upload process. My suspicion is still that there's something going on with permissions or some such but I know the hosting company are going to want me to be a bit more specific than that....

Justin