Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Strange behaviour of small allocator #5216

Closed
alyapunov opened this issue Jul 29, 2020 · 0 comments · Fixed by #5503
Closed

Strange behaviour of small allocator #5216

alyapunov opened this issue Jul 29, 2020 · 0 comments · Fixed by #5503
Assignees
@EvgenyMekhanik
Labels
bug Something isn't working crash
Milestone
2.5.3

Comments

@alyapunov
Copy link
Contributor

Tarantool version: any

The code

box.cfg{wal_mode='none', memtx_memory=2*1024*1024*1024, listen=3301, memtx_min_tuple_size=32, slab_alloc_factor=1.03}

s = box.schema.space.create('test')
_ = s:create_index('test')

function str(i) return string.rep('a', math.floor(256 * math.pow(1.03, i))) end
for i=1,276 do s:replace{i+200, str(i)} end
for i=1,274 do s:delete{i+200} end
collectgarbage('collect')
s:replace{200+277, str(275)}
s:delete{200+275}
collectgarbage('collect')
s:delete{200+276}
collectgarbage('collect')

breaks debug version of tarantool:

tarantool: /home/alyapunov/Work/tarantool/src/lib/small/small/small.c:291: mempool_find: Assertion `size >= upper_bound->objsize_min' failed.
Aborted (core dumped)

and puts release version to unexpected state (see mem_used):

tarantool> box.slab.stats()[#box.slab.stats()-2]
---
- mem_free: 17654224
  mem_used: 18446744073708674504
  item_count: 3069965951
  item_size: 877112
  slab_count: 1
  slab_size: 16777216
...
@alyapunov alyapunov added the bug Something isn't working label Jul 29, 2020
@kyukhin kyukhin added the crash label Aug 14, 2020
@kyukhin kyukhin added this to the 2.5.3 milestone Aug 14, 2020
EvgenyMekhanik added a commit that referenced this issue Nov 5, 2020
@EvgenyMekhanik
Fix bug in small allocator in case of small slab_alloc_factor
c3b2d70 
Change small allocator behavior and add test for #5216 issue

Closes #5216
@EvgenyMekhanik EvgenyMekhanik mentioned this issue Nov 5, 2020
Merged
EvgenyMekhanik added a commit that referenced this issue Nov 6, 2020
@EvgenyMekhanik
Fix bug in small allocator in case of small slab_alloc_factor
aa8bfb1 
Change small allocator behavior and add test for #5216 issue

Closes #5216
EvgenyMekhanik added a commit that referenced this issue Nov 9, 2020
@EvgenyMekhanik
Fix bug in small allocator in case of small slab_alloc_factor
ae21123 
Change small allocator behavior and add test for #5216 issue

Closes #5216
EvgenyMekhanik added a commit that referenced this issue Nov 12, 2020
@EvgenyMekhanik
Change small allocator behavior and add test for #5216 issue
62a431a 
Closes #5216
EvgenyMekhanik added a commit that referenced this issue Nov 12, 2020
@EvgenyMekhanik
Change small allocator behavior and add test for #5216 issue
996aa16 
Closes #5216
EvgenyMekhanik added a commit that referenced this issue Nov 12, 2020
@EvgenyMekhanik
Change small allocator behavior and add test for #5216 issue
8451a66 
Closes #5216
EvgenyMekhanik added a commit that referenced this issue Nov 12, 2020
@EvgenyMekhanik
Change small allocator behavior and add test for #5216 issue
65c058c 
Closes #5216
EvgenyMekhanik added a commit that referenced this issue Nov 12, 2020
@EvgenyMekhanik
Change small allocator behavior and add test for #5216 issue
fedc724 
Closes #5216
EvgenyMekhanik added a commit that referenced this issue Nov 13, 2020
@EvgenyMekhanik
Change small allocator behavior and add test for #5216 issue
d3429f5 
Closes #5216
EvgenyMekhanik added a commit that referenced this issue Nov 13, 2020
@EvgenyMekhanik
Change small allocator behavior, add test for #5216 issue
531ed49 
Change box/cfg test. Previously, if we set slab_alloc_factor > 2,
the small_alloc_create function quietly changed it to 2,
now I add assert slab_alloc_factor > 1.0 && slab_alloc_factor <= 2.

Closes #5216
EvgenyMekhanik added a commit that referenced this issue Nov 13, 2020
@EvgenyMekhanik
Change small allocator behavior, add test for #5216 issue
a4a71d4 
Change box/cfg test. Previously, if we set slab_alloc_factor > 2,
the small_alloc_create function quietly changed it to 2,
now I add assert slab_alloc_factor > 1.0 && slab_alloc_factor <= 2.

Closes #5216
EvgenyMekhanik added a commit that referenced this issue Nov 13, 2020
@EvgenyMekhanik
Change small allocator behavior, add test for #5216 issue
2527316 
Add a check that the slab_alloc_factor is in the range
(1.0, 2.0] and if it is not so changed it to 1.0001 or 2.0
respectively

Closes #5216
@kyukhin kyukhin added the teamC label Nov 18, 2020
EvgenyMekhanik added a commit that referenced this issue Dec 7, 2020
@EvgenyMekhanik
Change small allocator behavior, add test for #5216 issue
e9cd50f 
Add a check that the slab_alloc_factor is in the range
(1.0, 2.0] and if it is not so changed it to 1.0001 or 2.0
respectively

Closes #5216
EvgenyMekhanik added a commit that referenced this issue Dec 10, 2020
@EvgenyMekhanik
Change small allocator behavior, add test for #5216 issue
324dea1 
Add a check that the slab_alloc_factor is in the range
(1.0, 2.0] and if it is not so changed it to 1.0001 or 2.0
respectively

Closes #5216
EvgenyMekhanik added a commit that referenced this issue Dec 10, 2020
@EvgenyMekhanik
Change small allocator behavior, add test for #5216 issue
b4ba07e 
Add a check that the slab_alloc_factor is in the range
(1.0, 2.0] and if it is not so changed it to 1.0001 or 2.0
respectively

Closes #5216
EvgenyMekhanik added a commit that referenced this issue Dec 17, 2020
@EvgenyMekhanik
Change small allocator behavior, add test for #5216 issue
d8b8c74 
Add a check that the slab_alloc_factor is in the range
(1.0, 2.0] and if it is not so changed it to 1.0001 or 2.0
respectively

Closes #5216
EvgenyMekhanik added a commit that referenced this issue Dec 17, 2020
@EvgenyMekhanik
Change small allocator behavior, add test for #5216 issue
63be384 
Add a check that the slab_alloc_factor is in the range
(1.0, 2.0] and if it is not so changed it to 1.0001 or 2.0
respectively

Closes #5216
EvgenyMekhanik added a commit that referenced this issue Dec 21, 2020
@EvgenyMekhanik
memtx: change small allocator behavior
44a66e1 
Previously, in small allocator, memory pools
were allocated at the request, which in the case
of a small slab_alloc_factor led to use
pools with incorrect sizes. This patch changed
small allocator behavior, now we allocate pools
on the stage of allocator creation. Also we use
special function to find appropriate pool, which
is faster, then previous version with rbtree.
This change fixes #5216.

Also moved a check, that the slab_alloc_factor is in
the range (1.0, 2.0] from small allocator to memtx_engine.
If factor is not in range change it to 1.0001 or 2.0 respectively

Closes #5216
EvgenyMekhanik added a commit that referenced this issue Dec 22, 2020
@EvgenyMekhanik
memtx: change small allocator behavior
b19e4fd 
Previously, in small allocator, memory pools
were allocated at the request, which in the case
of a small slab_alloc_factor led to use
pools with incorrect sizes. This patch changed
small allocator behavior, now we allocate pools
on the stage of allocator creation. Also we use
special function to find appropriate pool, which
is faster, then previous version with rbtree.
This change fixes #5216.

Also moved a check, that the slab_alloc_factor is in
the range (1.0, 2.0] from small allocator to memtx_engine.
If factor is not in range change it to 1.0001 or 2.0 respectively

Closes #5216
EvgenyMekhanik added a commit that referenced this issue Dec 23, 2020
@EvgenyMekhanik
memtx: change small allocator behavior
a6b9b71 
Previously, in small allocator, memory pools
were allocated at the request, which in the case
of a small slab_alloc_factor led to use
pools with incorrect sizes. This patch changed
small allocator behavior, now we allocate pools
on the stage of allocator creation. Also we use
special function to find appropriate pool, which
is faster, then previous version with rbtree.
This change fixes #5216.

Also moved a check, that the slab_alloc_factor is in
the range (1.0, 2.0] from small allocator to memtx_engine.
If factor is not in range change it to 1.0001 or 2.0 respectively

Closes #5216
EvgenyMekhanik added a commit that referenced this issue Dec 24, 2020
@EvgenyMekhanik
memtx: change small allocator behavior
aec54d2 
Previously, in small allocator, memory pools
were allocated at the request, which in the case
of a small slab_alloc_factor led to use
pools with incorrect sizes. This patch changed
small allocator behavior, now we allocate pools
on the stage of allocator creation. Also we use
special function to find appropriate pool, which
is faster, then previous version with rbtree.
This change fixes #5216.

Also moved a check, that the slab_alloc_factor is in
the range (1.0, 2.0] from small allocator to memtx_engine.
If factor is not in range change it to 1.0001 or 2.0 respectively

Closes #5216
EvgenyMekhanik added a commit that referenced this issue Dec 24, 2020
@EvgenyMekhanik
memtx: change small allocator behavior
e94c1c7 
Previously, in small allocator, memory pools
were allocated at the request, which in the case
of a small slab_alloc_factor led to use
pools with incorrect sizes. This patch changed
small allocator behavior, now we allocate pools
on the stage of allocator creation. Also we use
special function to find appropriate pool, which
is faster, then previous version with rbtree.
This change fixes #5216.

Also moved a check, that the slab_alloc_factor is in
the range (1.0, 2.0] from small allocator to memtx_engine.
If factor is not in range change it to 1.0001 or 2.0 respectively

Closes #5216
EvgenyMekhanik added a commit that referenced this issue Dec 24, 2020
@EvgenyMekhanik
memtx: change small allocator behavior
0fad5d0 
Previously, in small allocator, memory pools
were allocated at the request, which in the case
of a small slab_alloc_factor led to use
pools with incorrect sizes. This patch changed
small allocator behavior, now we allocate pools
on the stage of allocator creation. Also we use
special function to find appropriate pool, which
is faster, then previous version with rbtree.
This change fixes #5216.

Also moved a check, that the slab_alloc_factor is in
the range (1.0, 2.0] from small allocator to memtx_engine.
If factor is not in range change it to 1.0001 or 2.0 respectively

Closes #5216
EvgenyMekhanik added a commit that referenced this issue Dec 24, 2020
@EvgenyMekhanik
memtx: change small allocator behavior
c92273f 
Previously, in small allocator, memory pools
were allocated at the request, which in the case
of a small slab_alloc_factor led to use
pools with incorrect sizes. This patch changed
small allocator behavior, now we allocate pools
on the stage of allocator creation. Also we use
special function to find appropriate pool, which
is faster, then previous version with rbtree.
This change fixes #5216.

Also moved a check, that the slab_alloc_factor is in
the range (1.0, 2.0] from small allocator to memtx_engine.
If factor is not in range change it to 1.0001 or 2.0 respectively

Closes #5216
EvgenyMekhanik added a commit that referenced this issue Dec 24, 2020
@EvgenyMekhanik
memtx: change small allocator behavior
1238e1d 
Previously, in small allocator, memory pools
were allocated at the request, which in the case
of a small slab_alloc_factor led to use
pools with incorrect sizes. This patch changed
small allocator behavior, now we allocate pools
on the stage of allocator creation. Also we use
special function to find appropriate pool, which
is faster, then previous version with rbtree.
This change fixes #5216.

Also moved a check, that the slab_alloc_factor is in
the range (1.0, 2.0] from small allocator to memtx_engine.
If factor is not in range change it to 1.0001 or 2.0 respectively

Closes #5216
EvgenyMekhanik added a commit that referenced this issue Dec 25, 2020
@EvgenyMekhanik
memtx: change small allocator behavior
bb93158 
Previously, in small allocator, memory pools
were allocated at the request, which in the case
of a small slab_alloc_factor led to use
pools with incorrect sizes. This patch changed
small allocator behavior, now we allocate pools
on the stage of allocator creation. Also we use
special function to find appropriate pool, which
is faster, then previous version with rbtree.
This change fixes #5216.

Also moved a check, that the slab_alloc_factor is in
the range (1.0, 2.0] from small allocator to memtx_engine.
If factor is not in range change it to 1.0001 or 2.0 respectively

Closes #5216
EvgenyMekhanik added a commit that referenced this issue Dec 25, 2020
@EvgenyMekhanik
memtx: change small allocator behavior
1789732 
Previously, in small allocator, memory pools
were allocated at the request, which in the case
of a small slab_alloc_factor led to use
pools with incorrect sizes. This patch changed
small allocator behavior, now we allocate pools
on the stage of allocator creation. Also we use
special function to find appropriate pool, which
is faster, then previous version with rbtree.
This change fixes #5216.

Also moved a check, that the slab_alloc_factor is in
the range (1.0, 2.0] from small allocator to memtx_engine.
If factor is not in range change it to 1.0001 or 2.0 respectively

Closes #5216
EvgenyMekhanik added a commit that referenced this issue Dec 28, 2020
@EvgenyMekhanik
memtx: change small allocator behavior
af4f7ba 
Previously, in small allocator, memory pools
were allocated at the request, which in the case
of a small slab_alloc_factor led to use
pools with incorrect sizes. This patch changed
small allocator behavior, now we allocate pools
on the stage of allocator creation. Also we use
special function to find appropriate pool, which
is faster, then previous version with rbtree.
This change fixes #5216.

Also moved a check, that the slab_alloc_factor is in
the range (1.0, 2.0] from small allocator to memtx_engine.
If factor is not in range change it to 1.0001 or 2.0 respectively

Closes #5216
EvgenyMekhanik added a commit that referenced this issue Dec 29, 2020
@EvgenyMekhanik
memtx: change small allocator behavior
282854a 
Previously, in small allocator, memory pools
were allocated at the request, which in the case
of a small slab_alloc_factor led to use
pools with incorrect sizes. This patch changed
small allocator behavior, now we allocate pools
on the stage of allocator creation. Also we use
special function to find appropriate pool, which
is faster, then previous version with rbtree.
This change fixes #5216.

Also moved a check, that the slab_alloc_factor is in
the range (1.0, 2.0] from small allocator to memtx_engine.
If factor is not in range change it to 1.0001 or 2.0 respectively

Closes #5216
kyukhin pushed a commit that referenced this issue Dec 29, 2020
@EvgenyMekhanik @kyukhin
memtx: change small allocator behavior
4d175bf 
Previously, in small allocator, memory pools
were allocated at the request, which in the case
of a small slab_alloc_factor led to use
pools with incorrect sizes. This patch changed
small allocator behavior, now we allocate pools
on the stage of allocator creation. Also we use
special function to find appropriate pool, which
is faster, then previous version with rbtree.
This change fixes #5216.

Also moved a check, that the slab_alloc_factor is in
the range (1.0, 2.0] from small allocator to memtx_engine.
If factor is not in range change it to 1.0001 or 2.0 respectively

Closes #5216
EvgenyMekhanik added a commit that referenced this issue Dec 29, 2020
@EvgenyMekhanik
memtx: change small allocator behavior
1295e04 
Previously, in small allocator, memory pools
were allocated at the request, which in the case
of a small slab_alloc_factor led to use
pools with incorrect sizes. This patch changed
small allocator behavior, now we allocate pools
on the stage of allocator creation. Also we use
special function to find appropriate pool, which
is faster, then previous version with rbtree.
This change fixes #5216.

Also moved a check, that the slab_alloc_factor is in
the range (1.0, 2.0] from small allocator to memtx_engine.
If factor is not in range change it to 1.0001 or 2.0 respectively

Closes #5216
kyukhin pushed a commit that referenced this issue Dec 29, 2020
@EvgenyMekhanik @kyukhin
memtx: change small allocator behavior
ad2f01d 
Previously, in small allocator, memory pools
were allocated at the request, which in the case
of a small slab_alloc_factor led to use
pools with incorrect sizes. This patch changed
small allocator behavior, now we allocate pools
on the stage of allocator creation. Also we use
special function to find appropriate pool, which
is faster, then previous version with rbtree.
This change fixes #5216.

Also moved a check, that the slab_alloc_factor is in
the range (1.0, 2.0] from small allocator to memtx_engine.
If factor is not in range change it to 1.0001 or 2.0 respectively

Closes #5216

(cherry picked from commit 4d175bf)
kyukhin pushed a commit that referenced this issue Dec 29, 2020
@EvgenyMekhanik @kyukhin
memtx: change small allocator behavior
47e12c7 
Previously, in small allocator, memory pools
were allocated at the request, which in the case
of a small slab_alloc_factor led to use
pools with incorrect sizes. This patch changed
small allocator behavior, now we allocate pools
on the stage of allocator creation. Also we use
special function to find appropriate pool, which
is faster, then previous version with rbtree.
This change fixes #5216.

Also moved a check, that the slab_alloc_factor is in
the range (1.0, 2.0] from small allocator to memtx_engine.
If factor is not in range change it to 1.0001 or 2.0 respectively

Closes #5216

(cherry picked from commit 4d175bf)
kyukhin pushed a commit that referenced this issue Dec 29, 2020
@EvgenyMekhanik @kyukhin
memtx: change small allocator behavior
6997264 
Previously, in small allocator, memory pools
were allocated at the request, which in the case
of a small slab_alloc_factor led to use
pools with incorrect sizes. This patch changed
small allocator behavior, now we allocate pools
on the stage of allocator creation. Also we use
special function to find appropriate pool, which
is faster, then previous version with rbtree.
This change fixes #5216.

Also moved a check, that the slab_alloc_factor is in
the range (1.0, 2.0] from small allocator to memtx_engine.
If factor is not in range change it to 1.0001 or 2.0 respectively

Closes #5216
@no1seman no1seman mentioned this issue Dec 19, 2021
Closed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@EvgenyMekhanik EvgenyMekhanik

Labels
bug Something isn't working crash
Projects
None yet
Milestone
2.5.3
Development

Successfully merging a pull request may close this issue.

Fix bug in small allocator in case of small slab_alloc_factor tarantool/tarantool
3 participants
@alyapunov @kyukhin @EvgenyMekhanik