Opened 12 years ago

Closed 22 months ago

Last modified 9 months ago

#3566 closed bug (fixed)

Data corruption when compiling under hybrid

Reported by: Adek336 Owned by: axeld
Priority: normal Milestone: R1/beta2
Component: File Systems/BFS Version: R1/pre-alpha1
Keywords: Cc:
Blocked By: Blocking:
Platform: All


I've been compiling some userland applications a lot recently. I have seen

under vmware, hybrid, two BFS partitions, code is being compiled on the non-boot partition, suddenly the build process ends with an error like garbage in config.status, syntax error in .c file, lack of separator in Makefile. These error always turned out to be caused by corruption in the mentioned files. This happened a lot in a revision circa hrev29330, now under circa hrev29500 I didn't see it but perhaps it's just a matter of chance. Perhaps related to #3347 which happened even more often than data corruption.

under real hw, hrev29503, hybrid, two partitions, compiling on non-boot partition, I've seen this happen once: compiling compiling then boom restart, after booting up many files are corrupt, even sources which were compiled an hour earlier didn't want to compile again due to garbage in files, some stuff in /boot/common/bin and /boot/common/lib didn't work because of elf header errors like if they were corrupted.

Attachments (3)

corruption.txt (6.7 KB ) - added by jessicah 6 years ago.
diff of the sections of a tarball that got corrupted
checkfs-output.txt (4.5 KB ) - added by jessicah 6 years ago.
Weird checkfs output
syslog-excerpt.txt (1.4 KB ) - added by jessicah 6 years ago.
some of the syslog whilst running checkfs

Download all attachments as: .zip

Change History (9)

comment:1 by diver, 9 years ago

Component: - GeneralFile Systems/BFS

comment:2 by bonefish, 7 years ago

Has this issue still been observed recently?

by jessicah, 6 years ago

Attachment: corruption.txt added

diff of the sections of a tarball that got corrupted

comment:3 by pulkomandy, 6 years ago

I have seen corruption happen, usually on the file while building webkit. Garbage would be found in the middle of the file. It seemed to be a cache problem, because rebooting the machine would solve it.

I haven't seen that issue again since BeGeistert, but I've not worked on WebKit as much since then. I'll keep looking for it.

comment:4 by jessicah, 6 years ago

This just happened to me today, after compiling GHC multiple times with haikuporter, trying to iron out some bugs.

Running checkfs multiple times, seems to find a new corrupted directory entry each run now.

On hrev48310 gcc2hybrid.

Last edited 6 years ago by jessicah (previous) (diff)

by jessicah, 6 years ago

Attachment: checkfs-output.txt added

Weird checkfs output

by jessicah, 6 years ago

Attachment: syslog-excerpt.txt added

some of the syslog whilst running checkfs

comment:5 by waddlesplash, 22 months ago

Resolution: fixed
Status: newclosed

Data corruption has not been seen in some time; any still-extant instances of it should be a new ticket.

comment:6 by nielx, 9 months ago

Milestone: R1R1/beta2

Assign tickets with status=closed and resolution=fixed within the R1/beta2 development window to the R1/beta2 Milestone

Note: See TracTickets for help on using tickets.