标签归档：pg

ceph

ceph pg 计算方法 python

2019 年 1 月 10 日 admin 发表回复

import math

pow_threshold = 0.25

def nearest_pow2(size):
    val = math.pow(2,round(math.log(size)/math.log(2)))
    if val < size * (1-pow_threshold):
        val *=2
    return val

def update_pg_count(osd_num,size,pg_osd=100,pre_data=0.1):

    min_val = nearest_pow2(math.floor(osd_num/size)+1)

    if min_val < osd_num:
        min_val *= 2
    calc_val = nearest_pow2(math.floor(pg_osd*osd_num*pre_data)/(100*size))

    if min_val > calc_val:
        return min_val
    else:
        return calc_val


print(int(update_pg_count(100,3,100,0.1)))

import math

pow_threshold = 0.25

def nearest_pow2(size):

val = math.pow(2,round(math.log(size)/math.log(2)))

if val < size * (1-pow_threshold):

val *=2

return val

def update_pg_count(osd_num,size,pg_osd=100,pre_data=0.1):

min_val = nearest_pow2(math.floor(osd_num/size)+1)

if min_val < osd_num:

min_val *= 2

calc_val = nearest_pow2(math.floor(pg_osd*osd_num*pre_data)/(100*size))

if min_val > calc_val:

return min_val

else:

return calc_val

print(int(update_pg_count(100,3,100,0.1)))

ceph

ceph的pool的pg修改值的计算

2017 年 5 月 3 日 admin 发表回复

pg的修改
当扩大pg num的时候，有时候会遇到报错：

ceph osd pool set rbd pg_num 4096
 Error E2BIG: specified pg_num 3500 is too large (creating 4096 new PGs \
 on ~64 OSDs exceeds per-OSD max of 32)

ceph osd pool set rbd pg_num 4096

Error E2BIG: specified pg_num 3500 is too large (creating 4096 new PGs \

on ~64 OSDs exceeds per-OSD max of 32)

限制pg spliting的参数来源于mon_osd_max_split_count value。
查看配置文件

#ceph daemon mon.ipmi1151 config get mon_osd_max_split_count
{
    "mon_osd_max_split_count": "32"
}

#ceph daemon mon.ipmi1151 config get mon_osd_max_split_count

{

"mon_osd_max_split_count": "32"

}

计算脚本：

max_inc=`ceph daemon mon.ipmi1151 config get mon_osd_max_split_count 2>&1 \
  | tr -d '\n ' | sed 's/.*"\([[:digit:]]\+\)".*/\1/'`
pg_num=`ceph osd pool get wtest pg_num | cut -f2 -d: | tr -d ' '`
echo "current pg_num value: $pg_num, max increment: $max_inc"
osd_num=`ceph osd ls |wc -l`
next_pg_num="$(($pg_num+$(($max_inc * $osd_num))))"
echo "allowed increment of pg_num: $next_pg_num"

max_inc=`ceph daemon mon.ipmi1151 config get mon_osd_max_split_count 2>&1 \

| tr -d '\n ' | sed 's/.*"$[[:digit:]]\+$".*/\1/'`

pg_num=`ceph osd pool get wtest pg_num | cut -f2 -d: | tr -d ' '`

echo "current pg_num value: $pg_num, max increment: $max_inc"

osd_num=`ceph osd ls |wc -l`

next_pg_num="$(($pg_num+$(($max_inc * $osd_num))))"

echo "allowed increment of pg_num: $next_pg_num"

结果输出

current pg_num value: 512, max increment: 32
allowed increment of pg_num: 800

1 2	current pg_num value: 512, max increment: 32 allowed increment of pg_num: 800

block_size的计算

block_size = mon_osd_max_split_count * n_osds.

( Target PGs per OSD ) x ( OSD # ) x ( %Data )
-------------------------------------------------
                     ( Size )

block_size = mon_osd_max_split_count * n_osds.

( Target PGs per OSD ) x ( OSD # ) x ( %Data )

-------------------------------------------------

( Size )

If the value of the above calculation is less than the value of ( OSD# ) / ( Size ), then the value is updated to the value of ( OSD# ) / ( Size ). This is to ensure even load / data distribution by allocating at least one Primary or Secondary PG to every OSD for every Pool.
The output value is then rounded to the nearest power of 2.
Tip: The nearest power of 2 provides a marginal improvement in efficiency of the CRUSH algorithm.
If the nearest power of 2 is more than 25% below the original value, the next higher power of 2 is used.

ceph

pg与osd的对应关系查找

2017 年 4 月 17 日 admin 发表回复

利用pg在pool以及osd的分布关系可以更好的分析pg的分布是否均匀

 pool:      0    1    2    3    4    5    6    7    8    9   10   11   12   13   14   15   16   17| SUM 
------------------------------------------------------------------------------------------------------- 
osd.0      29    9   13   12   10   14    9   14    7   13   10    9   11   18  355  329  305   49| 1216 
osd.1      25   11    6   11   16   10    7    7   13   11    9    8   15   14  316  369  339   40| 1227 
osd.2      23   10   12   10   16   13    8   11   14   10   10   11   14   23  344  339  340   37| 1245 
osd.3      22    8    8   10   10    9    9   12   12    9   12    9    8   21  325  337  361   40| 1222 
osd.4      13   15   11   10   12    9   14    6   13   10   10   14   13   25  344  358  358   39| 1274 
osd.5      13    9   11    9    8   12   15   13   10   15    9   12    7   24  339  327  355   46| 1234 
osd.6      25   13   10   10   11    7    8   10    8    8   13    6    9   23  322  337  349   46| 1215 
osd.7      16    9   10   12    5   12   16   11   10   14    9   15    9   18  358  348  335   45| 1252 
osd.8      26   12   15   12    8   10   10   12    9    6   14   12   10   26  369  328  330   42| 1251 
------------------------------------------------------------------------------------------------------- 
SUM:      192   96   96   96   96   96   96   96   96   96   96   96   96  192 3072 3072 3072  384 
MAX:       29   15   15   12   16   14   16   14   14   15   14   15   15   26  369  369  361   49 
MAX osd:   .0   .4   .8   .0   .1   .0   .7   .0   .2   .5   .8   .7   .1   .8   .8   .1   .3   .0 
osd Sum:    9    9    9    9    9    9    9    9    9    9    9    9    9    9    9    9    9    9 
MIN osd:   .5   .3   .1   .5   .7   .6   .1   .4   .0   .8   .1   .6   .5   .1   .1   .5   .0   .2

pool: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17| SUM

-------------------------------------------------------------------------------------------------------

osd.0 29 9 13 12 10 14 9 14 7 13 10 9 11 18 355 329 305 49| 1216

osd.1 25 11 6 11 16 10 7 7 13 11 9 8 15 14 316 369 339 40| 1227

osd.2 23 10 12 10 16 13 8 11 14 10 10 11 14 23 344 339 340 37| 1245

osd.3 22 8 8 10 10 9 9 12 12 9 12 9 8 21 325 337 361 40| 1222

osd.4 13 15 11 10 12 9 14 6 13 10 10 14 13 25 344 358 358 39| 1274

osd.5 13 9 11 9 8 12 15 13 10 15 9 12 7 24 339 327 355 46| 1234

osd.6 25 13 10 10 11 7 8 10 8 8 13 6 9 23 322 337 349 46| 1215

osd.7 16 9 10 12 5 12 16 11 10 14 9 15 9 18 358 348 335 45| 1252

osd.8 26 12 15 12 8 10 10 12 9 6 14 12 10 26 369 328 330 42| 1251

-------------------------------------------------------------------------------------------------------

SUM: 192 96 96 96 96 96 96 96 96 96 96 96 96 192 3072 3072 3072 384

MAX: 29 15 15 12 16 14 16 14 14 15 14 15 15 26 369 369 361 49

MAX osd: .0 .4 .8 .0 .1 .0 .7 .0 .2 .5 .8 .7 .1 .8 .8 .1 .3 .0

osd Sum: 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9

MIN osd: .5 .3 .1 .5 .7 .6 .1 .4 .0 .8 .1 .6 .5 .1 .1 .5 .0 .2

首先获取pg的query数据。利用Ansible模块获取数据。数据如下：

    "ceph_facts": {
        "update_pg_state": {
            "osds": {
                "0": {
                    "0": {
                        "pgsize": 29
                    }, 
                    "1": {
                        "pgsize": 9
                    }, 
                    "10": {
                        "pgsize": 10
                    }, 
                    "11": {
                        "pgsize": 9
...
                    "pgsum": 96
                }
            }
        }
    }

"ceph_facts": {

"update_pg_state": {

"osds": {

"0": {

"pgsize": 29

"1": {

"pgsize": 9

"10": {

"pgsize": 10

"11": {

"pgsize": 9

...

"pgsum": 96

}

利用python分析该数据：

def ceph_pg_parser(stat_dic):
    _str = ""
    pg_stat = stat_dic.get('ceph_facts')
    pool_pgs = pg_stat.get('update_pg_state',[]).get('pools',[])
    osd_pgs = pg_stat.get('update_pg_state',[]).get('osds',[])
    if len(pool_pgs) < 0 and len(osd_pgs) < 0: return
    _str += "pool%-3s" % ":"
    for pg in sorted(pool_pgs,key=int):
       _str += "%5s" %  pg
    _str += "|%5s" % "SUM\n"
    _str += "--------" + "-----" * len(pool_pgs) + "-----\n"
    for osdid in sorted(osd_pgs,key=int):
       _str += "osd.%-4s" % osdid
       for pg in sorted(pool_pgs,key=int):
          if pool_pgs[pg].has_key(osdid):
             _str += "%5s" % pool_pgs[pg][osdid]['pgsize']
          else:
             _str += "%5s" % 0
       _str += "|%5s\n" % osd_pgs[osdid]['pgsum']
    _str += "--------" + "-----" * len(pool_pgs) + "-----\n"
    _str += "SUM:    "
    for pg in sorted(pool_pgs,key=int):
      _str += "%5s" % pool_pgs[pg]['pgsum']
    _str += "\n"
    _str += "MAX:    "
    for pg in sorted(pool_pgs,key=int):
      _str += "%5s" % pool_pgs[pg]['max']
    _str += "\n"
    _str += "MAX osd:"
    for pg in sorted(pool_pgs,key=int):
      _str += "%5s" % ("." + str(pool_pgs[pg]['max_osd_id']))
    _str += "\n"
    _str += "OSD Sum:"
    for pg in sorted(pool_pgs,key=int):
      _str += "%5d" % (len(pool_pgs[pg])-int(5))
    _str += "\n"
    _str += "MIN osd:"
    for pg in sorted(pool_pgs,key=int):
      _str += "%5s" % ("." + str(pool_pgs[pg]['min_osd_id']))
    _str += "\n"
    return _str

def ceph_pg_parser(stat_dic):

_str = ""

pg_stat = stat_dic.get('ceph_facts')

pool_pgs = pg_stat.get('update_pg_state',[]).get('pools',[])

osd_pgs = pg_stat.get('update_pg_state',[]).get('osds',[])

if len(pool_pgs) < 0 and len(osd_pgs) < 0: return

_str += "pool%-3s" % ":"

for pg in sorted(pool_pgs,key=int):

_str += "%5s" % pg

_str += "|%5s" % "SUM\n"

_str += "--------" + "-----" * len(pool_pgs) + "-----\n"

for osdid in sorted(osd_pgs,key=int):

_str += "osd.%-4s" % osdid

for pg in sorted(pool_pgs,key=int):

if pool_pgs[pg].has_key(osdid):

_str += "%5s" % pool_pgs[pg][osdid]['pgsize']

else:

_str += "%5s" % 0

_str += "|%5s\n" % osd_pgs[osdid]['pgsum']

_str += "--------" + "-----" * len(pool_pgs) + "-----\n"

_str += "SUM: "

for pg in sorted(pool_pgs,key=int):

_str += "%5s" % pool_pgs[pg]['pgsum']

_str += "\n"

_str += "MAX: "

for pg in sorted(pool_pgs,key=int):

_str += "%5s" % pool_pgs[pg]['max']

_str += "\n"

_str += "MAX osd:"

for pg in sorted(pool_pgs,key=int):

_str += "%5s" % ("." + str(pool_pgs[pg]['max_osd_id']))

_str += "\n"

_str += "OSD Sum:"

for pg in sorted(pool_pgs,key=int):

_str += "%5d" % (len(pool_pgs[pg])-int(5))

_str += "\n"

_str += "MIN osd:"

for pg in sorted(pool_pgs,key=int):

_str += "%5s" % ("." + str(pool_pgs[pg]['min_osd_id']))

_str += "\n"

return _str

SELinux+

标签归档：pg

ceph pg 计算方法 python

ceph的pool的pg修改值的计算

pg与osd的对应关系查找

基础分享安全

2019年一月
一	二	三	四	五	六	日
« 十
	1	2	3	4	5	6
7	8	9	10	11	12	13
14	15	16	17	18	19	20
21	22	23	24	25	26	27
28	29	30	31

基础 分享 安全

基础分享安全