[Gluster-devel] spurious split brain (was: Re: glusterfs-3.4.0qa2 released)

[Emmanuel Dreyfus]

Emmanuel Dreyfus <manu at netbsd.org> wrote:

> Here is the log. My first hunch was wrong, the split brain does not
> seems to come from a disconnect/reconnect cycle:
> http://ftp.espci.fr/shadow/manu/gfs340qa2-spurious-eio.log

I collected the FUSE trace for operation on the offending file (which is
different from the previos trace above)

1352468511.995130403 LOOKUP "gcc" path = "postreload-gcse.o" 
  error = 0 elapsed = 0.778808760 
1352468512.774222289 OPEN "postreload-gcse.o"  
  error = 0 elapsed = 0.000132672 
1352468513.030707629 RELEASE "postreload-gcse.o" 
  error = 0 elapsed = 0.000066234 
1352468516.279496048 LOOKUP "gcc" path = "postreload-gcse.o" 
  error = 0 elapsed = 3.456378598 
1352468523.485015107 LOOKUP "gcc" path = "postreload-gcse.o" 
  error = 0 elapsed = 0.688687475 
1352468524.173978661 OPEN "postreload-gcse.o"  
  error = 0 elapsed = 0.000123981 
1352468524.174466835 READ "postreload-gcse.o"  
  error = EIO elapsed = 0.000403626

Is there anything shocking here?

-- 
Emmanuel Dreyfus
http://hcpnet.free.fr/pubz
manu at netbsd.org