[LU-7318] OUT: dynamic reply buffer Created: 20/Oct/15 Updated: 24/Nov/15 Resolved: 24/Nov/15 |
|
| Status: | Resolved |
| Project: | Lustre |
| Component/s: | None |
| Affects Version/s: | Lustre 2.8.0 |
| Fix Version/s: | Lustre 2.8.0 |
| Type: | Bug | Priority: | Blocker |
| Reporter: | Alex Zhuravlev | Assignee: | Alex Zhuravlev |
| Resolution: | Fixed | Votes: | 0 |
| Labels: | None | ||
| Issue Links: |
|
||||||||||||||||
| Rank (Obsolete): | 9223372036854775807 | ||||||||||||||||
| Description |
|
osp_prep_update_req() allocates 4K as a reply buffer where 4K is hardcoded. instead buffer size should depend on the number of update in the request. similarly, the target should check whether reply buffer is big enough to store all the results. |
| Comments |
| Comment by Gerrit Updater [ 20/Oct/15 ] |
|
Alex Zhuravlev (alexey.zhuravlev@intel.com) uploaded a new patch: http://review.whamcloud.com/16889 |
| Comment by Di Wang [ 20/Oct/15 ] |
|
Actually, after we change OUT_UPDATE_REPLY_SIZE from 16384 to 4096 ( http://review.whamcloud.com/15899), this problem already bite us, because 4096 is certainly not enough for some big updates operation like sanity.sh 300k "test large striped directory". Simply run "MDSCOUNT=4 ONLY=300 sh sanity.sh" 2/3 times, will cause memory corruption. So this patch needs to get into 2.8. |
| Comment by James A Simmons [ 20/Oct/15 ] |
|
Agree. I also have seen sanity test 300k take out my nodes. |
| Comment by John Hammond [ 13/Nov/15 ] |
|
Di, Why wasn't the stupid test disabled 4+ weeks ago? |
| Comment by Gerrit Updater [ 24/Nov/15 ] |
|
Oleg Drokin (oleg.drokin@intel.com) merged in patch http://review.whamcloud.com/16889/ |
| Comment by Joseph Gmitter (Inactive) [ 24/Nov/15 ] |
|
Landed for 2.8 |