Bug 8524 - Files Saves on 3.5.11 randomly fail due to non-released pseudolocks.
Files Saves on 3.5.11 randomly fail due to non-released pseudolocks.
Status: NEW
Product: Samba 3.5
Classification: Unclassified
Component: File services
x64 All
: P5 major
: ---
Assigned To: Volker Lendecke
Samba QA Contact
Depends on:
  Show dependency treegraph
Reported: 2011-10-13 16:35 UTC by Linda Walsh
Modified: 2011-10-13 16:35 UTC (History)
0 users

See Also:


Note You need to log in before you can comment on or make changes to this bug.
Description Linda Walsh 2011-10-13 16:35:28 UTC
I'm not sure what triggers it, but it is more likely to happen on larger files or when the network is inexplicably slow.  

I'm not certain, but I think it only happens with programs that
1) precreate the intended target name and open it with a write lock.
   1a) I've seen evidence, that at least in some cases, the file is
       created as a sparse file by that would be the size of the final
       file (in one case of a system crash during it copying the profile 
       to the server, where it uses tmp & rename, several files in my profile
       were filled with zeros -- of the same length as 'tmp files' that were'
       in the same directory (that were the right files for the same-sized 
       null-filled files).
2) write to a tmp file during the active download or copy.
3) then rename the tmp to the final copy -- which fails!...  errors come up
up like permission denied, disk no longer available, 'unknown failure writing
to disk...etc...

Thing that is odd -- is there is some other factor that figures into this, 
because I can save the same file to the same dir sometimes and have it work.
It depends on network conditions.  Sometimes the network seems near normal and
writes go at more normal speeds, but more often, of late, the speeds have been
slower, and when it is slow, is when it is more likely to fail..

Maybe I'm confused, but I thought this was only a problem in 3.6.   I mentioned at the time, that I thought it sounded similar to a bug someone ELSE was having (same type of problem) in 3.5.11, but was told it was only in 3.6.   So why am I seeing a similar symptom?