Bug 767948
| Summary: | [7235e5b1af090ffc9d87ac59daadf7926433b495] fsx segfaults for large number of operations on fusemount | ||
|---|---|---|---|
| Product: | [Community] GlusterFS | Reporter: | Rahul C S <rahulcs> |
| Component: | io-cache | Assignee: | Raghavendra G <rgowdapp> |
| Status: | CLOSED CURRENTRELEASE | QA Contact: | |
| Severity: | medium | Docs Contact: | |
| Priority: | medium | ||
| Version: | mainline | CC: | gluster-bugs, kaushal, rfortier |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | glusterfs-3.4.0 | Doc Type: | Bug Fix |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2013-07-24 17:38:32 UTC | Type: | --- |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | 1f3a0dd4742a2fcd3215aee4a5e22125d7ea4f4d | Category: | --- |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
| Bug Depends On: | |||
| Bug Blocks: | 817967 | ||
|
Description
Rahul C S
2011-12-15 10:53:43 UTC
The segmentation fault here is a problem of fsx not gluster. The segfault occurs while printing the log to the logfile. But this segfault is hit because one of the reads done by fsx, returns with a size smaller than what was requested. Testing this further with a small program which tries to read the the same data that fsx tried and failed. When I run this program on the file on the mount point, read returns a size lesser than the size requested, as in fsx. But running the program on the file on the brick, read returns the correct size. Some more observations after further testing. The short read in fsx seems to happen at holes in the file, mostly those holes which are at the end of the file. This problem goes away when disabling the io-cache translator. This behavior is similar to https://bugzilla.redhat.com/show_bug.cgi?id=782286 Put in an incorrect bug link above. The correct bug is https://bugzilla.redhat.com/show_bug.cgi?id=788962 CHANGE: http://review.gluster.com/2981 (performance/io-cache: destroy the pages which survived in __ioc_page_destroy because of non NULL waitq in ioc_fault_cbk.) merged in master by Vijay Bellur (vijay) CHANGE: http://review.gluster.com/2982 (performance/read-ahead: destroy pages which survived from being flushed in flush_region because of frames waiting on them, after those frames are served.) merged in master by Vijay Bellur (vijay) root@Dagobah:/data/mounts/fuse# /opt/qa/tools/fsx -R -W -N 500000 file mapped writes DISABLED truncating to largest ever: 0x32740 truncating to largest ever: 0x39212 truncating to largest ever: 0x3bae9 truncating to largest ever: 0x3c1e3 truncating to largest ever: 0x3d1cd truncating to largest ever: 0x3e8b8 truncating to largest ever: 0x3ed14 truncating to largest ever: 0x3f9c2 truncating to largest ever: 0x3ff9f truncating to largest ever: 0x3ffa9 truncating to largest ever: 0x3ffb0 truncating to largest ever: 0x3ffce truncating to largest ever: 0x3ffe0 truncating to largest ever: 0x3fff1 skipping zero size read truncating to largest ever: 0x3fff4 skipping zero size read skipping zero size write skipping zero size write skipping zero size read skipping zero size read skipping zero size read skipping zero size read truncating to largest ever: 0x3fffe skipping zero size read All operations completed A-OK! Works as expected with latest git head. |