Jonathan Boutelle wrote:
> The server and all clients are running Norton Antivirus with Autoprotect
> enabled. So that is a potential explanation for my problem. Not about to
> disable my virus scanning however: I get multiple copies of the netski
> virus etc mailed to me every day!
> J
>
> At 06:32 AM 6/23/2004, jal0 wrote:
>
>> Just a question...
>>
>> Are those of you experiencing this problem running any antivirus
>> software?
>>
>> I've seen a similar thing happening in Microsoft Visual Studios. When
>> initiating a build, I get an "unable to access file. File may be
>> missing or in use". If I immediately initiate another build it goes
>> through with no problems.
>>
>> The last time I saw this error message I went to explorer and the file
>> in question had been renamed to some random file name with a .tmp
>> extension .I opened it up in a txt editor as it _was_ the same file...
>> only something had decided to rename it at exactly the same time msvc
>> was trying to compile it.
>>
>> By hitting the 'build' button, msvc re-saves all files automatically
>> before rebuilding so no damage was done. Although once I _didn't_ do
>> an immediate rebuild, but closed the application. The next time I open
>> the project, the file had disappeared. That really sucked. Lucking I
>> had a backup in my subversion repository :)
>>
>> Anyway, I did some searching on google and found out that other had
>> experienced the same issue and almost all of them had some form of
>> 'Real Time Protection' or antivirus software installed. And the reason
>> for the problems stemmed from the antivirus software trying to scan
>> the file when a application first accesses it. If the timing is right
>> the conflict arises.
>>
>> I've also seen this issue manifest itself when trying to delete a
>> file. Windows aborts the operation with an error stating it cant
>> delete the file due to it's being used by another process... the
>> antivirus software.
>>
>> Anyway, maybe this is the problem some of you are seeing this happen
>> sporadically. Try disabling the antivirus software for a while ( when
>> you not surfing the net of course).
>>
>> Hope this helps. This has been a bugger of a problem for me and I
>> haven't been able to find a solution yet.
>>
>> Cheerz,
>> Jal0
>>
>> -----Original Message-----
>> From: Jonathan Boutelle [mailto:jon@uzanto.com]
>> Sent: Friday, June 18, 2004 9:12 PM
>> To: Branko Čibej
>> Cc: users@subversion.tigris.org
>> Subject: Re: scary "file not found" failures
>>
>> If all someone is doing is constantly updating a subsection of the tree,
>> and occasionally individual files (via.Tortoise SVN), where would the
>> potential confusion come from that would cause intermittent failure in
>> this
>> manner?
>>
>> Is there some mistake you can make during setup that would cause
>> intermittent failure at runtime? (Maybe something to do with the file
>> system of the server being FAT32 instead of NTFS or something?). Because
>> frankly I'm not doing anything complicated enough to mess up when the
>> error
>> occurs: I'm just right-clicking and selecting update.
>>
>> I will try to isolate when this bug occurs: it IS happening
>> (sporadically)
>> on multiple machines in my network.
>>
>> Regards,
>> Jon
>>
>> At 05:46 PM 6/18/2004, you wrote:
>> >Jonathan Boutelle wrote:
>> >
>> > > Oh - boy. Well at least I'm not alone. Yeah, from the reading that
>> > > sounds like exactly what I've got.
>> > > Am I correct that efforts to isolate this bug have foundered (the bug
>> > > is "closed" in the database)?
>> >
>> >Yes, I closed the bug because I am unable to reproduce it, and nobody
>> >else has yet found a recipe that consistently produces this error. So I
>> >suspect it's user confusion, compounded by a not-as-clear-as-could be
>> >error message.
>> >
>> > > BTW, updating at a "higher" level (the root, instead of the "rumm"
>> > > directory only) fixed the problem (which doesn't make any sense, but
>> > > there you go).
>> >
>> >Well, _that_ sounds like it could be a bug. I've seen something similar
>> >happen in other cases on Windows, but again never found a consistently
>> >reproducible case. If you could figure out exactly what happened (in
>> >your working copy and on the server), it would be a great help. If you
>> >do find a recipe, just attach it to issue 1862, and I'll pick it up from
>> >there.
>> >
>> >--
>> >Brane Èibej <brane_at_xbc.nu> http://www.xbc.nu/brane/
>>
>> ----^-------^------^--------^-------^
>> Jon Boutelle
>> Principal, Uzanto Consulting
>> San Francisco, CA
>>
>> Phone: 510-708-9825
>> urls: www.uzanto.com
>> www.jonathanboutelle.com
>> ----^-------^------^--------^-------^
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
>> For additional commands, e-mail: users-help@subversion.tigris.org
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
>> For additional commands, e-mail: users-help@subversion.tigris.org
>
>
> ----^-------^------^--------^-------^
> Jon Boutelle
> Principal, Uzanto Consulting
> San Francisco, CA
>
> Phone: 510-708-9825
> urls: www.uzanto.com
> www.jonathanboutelle.com
> ----^-------^------^--------^-------^
>
We had similar "odd" problems with failures occurring at random points
during checkouts. Tracking it down we discovered that it was strictly
to do with the real-time scanning being done with Inoculate-It (CA). It
would occur randomly when but became reproducible when trying to check
out an entire full tree, but never occurred in the same location.
Shutting down the real-time scan eliminated the problem. We did set up
the corporate policy to exclude working directories from the real-time,
and effected scheduled scans of those when the impact would not be seen.
Since that type we have switched to Symantec AV and with it's real-time
enabled the problem has not been seen. Believe it is something about
the way the CA code is accessing the files during the scan process.
This occurred both on XP and 2000 machines.
Hope this helps.
-- Mark
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Wed Jun 23 15:03:09 2004