"dog vdi object location" PANIC when "node < x"
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
sheepdog |
Fix Committed
|
Undecided
|
Unassigned |
Bug Description
This problem is similar to the report already as below , but occurrence conditions may slightly different.
https:/
How to reproduce:
# sheep -v
Sheepdog daemon version 0.9.0_101_
# dog cluster format -c 4
Number of copies (4) is larger than number of nodes (3).
Are you sure you want to continue? [yes/no]: yes
using backend plain store
# dog node list
Id Host:Port V-Nodes Zone
0 172.16.2.103:7000 128 0
1 172.16.2.103:7001 128 1
2 172.16.2.103:7002 128 2
# dog vdi create test 1M
# dog vdi list
Name Id Size Used Shared Creation time VDI id Copies Tag Block Size Shift
test 0 1.0 MB 0.0 MB 0.0 MB 2015-03-02 15:59 7c2b25 4 22
# dog vdi object location test
Looking for the inode object 0x7c2b25 with 3 nodes
172.16.2.103:7000 has the object
172.16.2.103:7001 has the object
172.16.2.103:7002 has the object
According to sheepdog algorithm, the object should be located at:
PANIC: can't find a valid vnode
dog exits unexpectedly (Aborted).
dog.c:375: crash_handler
/lib64/
/lib64/
/lib64/
sheep.h:80: oid_to_vnodes
vdi.c:1172: vdi_object_location
common.c:280: do_generic_
dog.c:576: main
/lib64/
dog() [0x403238]
[root@dfs1103 sheepdog-0302]# sheep -v
Sheepdog daemon version 0.9.0_101_
summary: |
- "dog vdi object location" clushed when "node < x" + "dog vdi object location" PANIC when "node < x" |
Changed in sheepdog-project: | |
status: | New → Fix Committed |
I tested in my reproduce environment with version below.
# sheep -v gb08ee11_ dirty.
Sheepdog daemon version 0.9.0_106_
it works good to me.