Closed GoogleCodeExporter closed 8 years ago
GoogleCodeExporter
commented
8 years ago the attached code will show data corruption in the zram device:
after a few cycles you will get e.g.:
cycle 8:
536184 tmpmnt
62398464
tmpmnt/tmpfile zram0mnt/f differ: char 4097, line 64
warning: this code may lead to oom conditions and therefore crash your systemOriginal comment by fadb24bb...@drewag.de on 21 Jan 2011 at 5:34
Attachments:
GoogleCodeExporter
commented
8 years ago I can now reproduce the issue using your script. Thanks. - Nitin
Original comment by nitingupta910@gmail.com on 22 Jan 2011 at 1:17
GoogleCodeExporter
commented
8 years ago I just sync'ed code in the repository with the mainline version (as in 2.6.37).
It does not contain many of the scalability improvements present in the version
you tested but this mainline version is more stable (I hope!).
So, can you please pull again and give it a try?Original comment by nitingupta910@gmail.com on 22 Jan 2011 at 4:13
GoogleCodeExporter
commented
8 years ago Great! no more random crashes, running: chrome, thunderbird, win 7, win xp,
eclipse
OOT: But after a few minutes, disk activity seem increased make system
unresponsive. Maybe because of low disk cache.
free -m
total used free shared buffers cached
Mem: 1998 1982 15 0 0 89
-/+ buffers/cache: 1892 106
Swap: 1953 1326 626
swapon -s
Filename Type Size Used Priority
/dev/zram0 partition 999996 684132 100
/dev/zram1 partition 999996 685780 100
Original comment by superbiji on 23 Jan 2011 at 7:02
GoogleCodeExporter
commented
8 years ago You are having total zram disksizes (zram0 + zram1) nearly equal to amount of
RAM i.e. 2G. Even if compression ratio is nice 50%, it will seriously reduce
RAM space available for filesystem cache (aka pagecache). So, you can probably
try reducing zram sizes to say 256M each to see how it goes?
Ideally, to make balance between swapcache and pagecache, we need support for
pagecache compression too. Currently, there are two parallel efforts (zcache
and kztmem). Hopefully, these efforts will merge into a single codebase and
accepted in mainline.
Marking as fixed.Original comment by nitingupta910@gmail.com on 23 Jan 2011 at 11:57
Original issue reported on code.google.com by
superbijion 11 Jan 2011 at 8:30