Bug 1460919
Summary: | Dead lock hit due to wrong order of locking in open behind and hence leading to fuse mount and client node hung | ||
---|---|---|---|
Product: | [Red Hat Storage] Red Hat Gluster Storage | Reporter: | Nag Pavan Chilakam <nchilaka> |
Component: | open-behind | Assignee: | Susant Kumar Palai <spalai> |
Status: | CLOSED DEFERRED | QA Contact: | Rahul Hinduja <rhinduja> |
Severity: | high | Docs Contact: | |
Priority: | high | ||
Version: | rhgs-3.3 | CC: | amukherj, mchangir, rgowdapp, rhs-bugs, storage-qa-internal |
Target Milestone: | --- | Keywords: | ZStream |
Target Release: | --- | Flags: | ykaul:
needinfo-
ykaul: needinfo- ykaul: needinfo- |
Hardware: | Unspecified | ||
OS: | Unspecified | ||
Whiteboard: | |||
Fixed In Version: | Doc Type: | If docs needed, set a value | |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2019-12-31 07:21:08 UTC | Type: | Bug |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: |
Description
Nag Pavan Chilakam
2017-06-13 07:17:08 UTC
RPM and other test info at http://etherpad.corp.redhat.com/nchilaka-Discoverer-Accelerated-Testing https://docs.google.com/spreadsheets/d/1ik5nqPJB1wtoFjhkFpk2RPcGBuL76yQ-gaQ-WPR_atI/edit#gid=0 Two mount process were deadlocked and one of them had active inodelks on a brick. Both the deadlocks involved interaction between open-behind and itable dump code Deadlock1: statedump had acquired locks in the order: itable -> inode -> fd openbehind was trying to acquire locks in the order: fd -> itable (as part of inode_ref done after acquiring fd->lock) Deadlock2: statedump was trying to acquire lock: itable -> inode openbehind was trying to acquire lock: inode -> itable (it was trying to create an anonymous fd which involves doing an inode_ref after acquiring inode->lock) Both the deadlocks are present in upstream master. Is this bug addressed in upstream now? Or should we track this in upstream? Closing - it received no real activity for ages. Please re-open if relevant and will be taken care of. |