| 1 |
Duncan wrote: |
| 2 |
|
| 3 |
> If you're interested in more, check the last three weeks' or so LWN |
| 4 |
> (lwn.net) kernel pages. There's links to the LKML threads, as well as to |
| 5 |
> the Ubuntu bug that kicked it all off. (Some proprietaryware NVidia |
| 6 |
> driver folks were crashing frequently, and those who had chosen the new |
| 7 |
> Ubuntu ext4 option noticed their gnome settings files were zeroed out |
| 8 |
> after they rebooted. That particular bug is papered over for 2.6.29 and |
| 9 |
> generally fixed for 2.6.30, but there's still a smaller risk as long as |
| 10 |
> you use the default data=writeback, the 2.6.30 ext3 default as well. So |
| 11 |
> as far as is known, with .30, ext4 SHOULD be as stable as ext3 in the |
| 12 |
> same data= mode, but the fact that such things are still coming up means |
| 13 |
> nobody's sure there aren't more such things to come up, thus, people who |
| 14 |
> value their data will want to wait to at least 2.6.31 before trying ext4, |
| 15 |
> and will want to be sure they specifically set data=ordered for ext3 |
| 16 |
> before switching to 2.6.30.) |
| 17 |
> |
| 18 |
> |
| 19 |
|
| 20 |
|
| 21 |
Those issues alone shouldn't cost you valuable data or cause filesystem |
| 22 |
loss, especially since they were patched. |
| 23 |
Even when I was loosong files, I never lost any really important |
| 24 |
archive, just bits and pieces of files that machine has modified, |
| 25 |
created or overwritten in some way, which means config files, libraries |
| 26 |
and binaries ( after emerge for example). |
| 27 |
I have applied patches to gentoo-sources manually just after a have |
| 28 |
grep-ed them from 2.6.28 patch on www.kernel.org, and from then I never |
| 29 |
had an issue. And BTW I do have journal=writeback all the time. |
| 30 |
|
| 31 |
Here are my mount options from /etc/fstab: |
| 32 |
|
| 33 |
1. My machine: |
| 34 |
|
| 35 |
/dev/sda5 / ext4 |
| 36 |
rw,noatime,user_xattr,acl,journal_checksum,journal_async_commit,data=ordered,barrier=0,commit=120,errors=remount-ro,max_batch_time=60000,min_batch_time=5000 |
| 37 |
0 1 |
| 38 |
|
| 39 |
|
| 40 |
2. One server: |
| 41 |
|
| 42 |
/dev/md1 / ext4 |
| 43 |
rw,noatime,user_xattr,acl,journal_checksum,journal_async_commit,data=ordered,barrier=0,commit=120,errors=remount-ro |
| 44 |
0 1 |
| 45 |
|
| 46 |
3. Another server: |
| 47 |
|
| 48 |
/dev/md2 / ext4 |
| 49 |
rw,noatime,user_xattr,acl,journal_checksum,journal_async_commit,data=ordered,barrier=0,commit=120,errors=continue |
| 50 |
0 1 |
| 51 |
|
| 52 |
|
| 53 |
|
| 54 |
All this time I had 0 problems with it on any of the machines. |
Archives