[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: DRBD Primary/Primary






2013/7/4 "Артём Н." <artiom14@yandex.ru>
Конфиги DRBD и Pacemaker во вложении.

Split brain:
version: 8.3.11 (api:88/proto:86-96)
srcversion: F937DCB2E5D83C6CCE4A6C9
 0: cs:StandAlone ro:Primary/Unknown ds:UpToDate/DUnknown   r-----
    ns:2796 nr:1832 dw:4628 dr:124530 al:6 bm:0 lo:0 pe:0 ua:0 ap:0 ep:1 wo:f oos:0
 1: cs:StandAlone ro:Primary/Unknown ds:UpToDate/DUnknown   r-----
    ns:0 nr:0 dw:0 dr:784 al:0 bm:0 lo:0 pe:0 ua:0 ap:0 ep:1 wo:f oos:0

10: cs:StandAlone ro:Primary/Unknown ds:UpToDate/DUnknown   r-----
    ns:104696 nr:1136 dw:105832 dr:186265 al:82 bm:3 lo:0 pe:0 ua:0 ap:0 ep:1
wo:f oos:0


Периодически всё работает нормально:
  0:system_data  Connected  Primary/Primary UpToDate/UpToDate     C r-----
  1:vm_volumes   Connected  Primary/Primary UpToDate/UpToDate     C r-----
 10:repository   SyncTarget Primary/Primary Inconsistent/UpToDate C r-----
        [=========>..........] sync'ed: 54.2% (432504/933888)K
...
  0:system_data  Connected Primary/Primary UpToDate/UpToDate C r-----
/mnt/system ocfs2 90G  968M 90G  2%
  1:vm_volumes   Connected Primary/Primary UpToDate/UpToDate C r-----
 10:repository   Connected Primary/Primary UpToDate/UpToDate C r----- /mnt/repo
  ocfs2 250G 82G  169G 33%
...

 Master/Slave Set: ms_drbd_repo [p_drbd_repo]
     Masters: [ cluster-data-1 cluster-data-2 ]
 Master/Slave Set: ms_drbd_system_data [p_drbd_system_data]
     Masters: [ cluster-data-1 cluster-data-2 ]
 Master/Slave Set: ms_drbd_vm_volumes [p_drbd_vm_volumes]
     Masters: [ cluster-data-1 cluster-data-2 ]
 Clone Set: ce_ocfs2mgmt [g_ocfs2mgmt]
     Started: [ cluster-data-1 cluster-data-2 ]
 Clone Set: ce_mysql [p_mysql]
     Started: [ cluster-data-1 cluster-data-2 ]
 Clone Set: ce_system_fs [p_system_fs]
     Started: [ cluster-data-1 cluster-data-2 ]
 Clone Set: ce_rabbitmq [p_rabbitmq]
     Started: [ cluster-data-1 cluster-data-2 ]
 Clone Set: ce_repo_fs [p_repo_fs]
     Started: [ cluster-data-1 cluster-data-2 ]
 Clone Set: ce_data_ip [p_data_ip]
     Started: [ cluster-data-1 cluster-data-2 ]
 Clone Set: ce_webserver [p_webserver]
     Started: [ cluster-data-1 cluster-data-2 ]


В Primary/Primary заработало после перевода в Primay/Secondary и обратно:

master-max у ms_drbd_system_data был 1.

root@cluster-data-1:~# drbd-overview
  0:system_data  StandAlone Primary/Unknown UpToDate/DUnknown     r-----
  1:vm_volumes   SyncTarget Primary/Primary Inconsistent/UpToDate C      r-----
        [>....................] sync'ed:  1.7% (729260/741688)Mfinish: 1:44:23
speed: 119,200 (112,632) want: 1,000,001 K/sec
 10:repository   Connected  Primary/Primary UpToDate/UpToDate     C      r-----

root@cluster-data-1:~# drbd-overview
^C

root@cluster-data-1:~# crm resource stop ms_drbd_system_data
root@cluster-data-1:~# crm configure edit ms_drbd_system_data

root@cluster-data-1:~# drbd-overview
  0:system_data  StandAlone Primary/Unknown UpToDate/DUnknown     r-----
  1:vm_volumes   SyncTarget Primary/Primary Inconsistent/UpToDate C      r-----
        [>....................] sync'ed:  3.0% (719600/741688)Mfinish: 1:46:58
speed: 114,784 (113,096) want: 1,000,001 K/sec
 10:repository   Connected  Primary/Primary UpToDate/UpToDate     C      r-----

root@cluster-data-1:~# drbd-overview
  0:system_data  StandAlone Primary/Unknown UpToDate/DUnknown     r-----
  1:vm_volumes   SyncTarget Primary/Primary Inconsistent/UpToDate C      r-----
        [>....................] sync'ed:  3.3% (717228/741688)Mfinish: 1:45:44
speed: 115,744 (113,340) want: 1,000,001 K/sec
 10:repository   Connected  Primary/Primary UpToDate/UpToDate     C      r-----

root@cluster-data-1:~# crm resource manage ms_drbd_system_data
root@cluster-data-1:~# drbd-overview
  0:system_data  Unconfigured .               .                     . .
  1:vm_volumes   SyncTarget   Primary/Primary Inconsistent/UpToDate C r-----
        [>....................] sync'ed:  4.2% (710848/741688)Mfinish: 1:45:17
speed: 115,212 (113,600) want: 1,000,001 K/sec
 10:repository   Connected    Primary/Primary UpToDate/UpToDate     C r-----

root@cluster-data-1:~# crm resource start ms_drbd_system_data
root@cluster-data-1:~# drbd-overview
  0:system_data  WFBitMapS  Primary/Secondary UpToDate/Consistent   C r-----
  1:vm_volumes   SyncTarget Primary/Primary   Inconsistent/UpToDate C r-----
        [>....................] sync'ed:  4.8% (706760/741688)Mfinish: 1:41:44
speed: 118,540 (113,908) want: 1,000,001 K/sec
 10:repository   Connected  Primary/Primary   UpToDate/UpToDate     C r-----
root@cluster-data-1:~# drbd-overview
  0:system_data  Connected  Primary/Secondary UpToDate/UpToDate     C r-----
  1:vm_volumes   SyncTarget Primary/Primary   Inconsistent/UpToDate C r-----
        [>...................] sync'ed:  5.2% (703360/741688)Mfinish: 1:44:21
speed: 115,008 (113,764) want: 1,000,001 K/sec
 10:repository   Connected  Primary/Primary   UpToDate/UpToDate     C r-----

root@cluster-data-1:~# crm configure edit ms_drbd_system_data
root@cluster-data-1:~# crm configure show ms_drbd_system_data
ms ms_drbd_system_data p_drbd_system_data \
        meta notify="true" clone-max="2" master-max="2" target-role="Started"
is-managed="true"
root@cluster-data-1:~# drbd-overview
  0:system_data  Connected  Primary/Primary UpToDate/UpToDate     C r-----
  1:vm_volumes   SyncTarget Primary/Primary Inconsistent/UpToDate C r-----
        [>...................] sync'ed:  7.0% (690308/741688)Mfinish: 1:40:21
speed: 117,388 (114,128) want: 1,000,001 K/sec
 10:repository   Connected  Primary/Primary UpToDate/UpToDate     C r-----

Иногда получается так, что вообще не останавливается ресурс:
root@cluster-data-1:~# cat /proc/drbd
version: 8.3.11 (api:88/proto:86-96)
srcversion: F937DCB2E5D83C6CCE4A6C9
 0: cs:StandAlone ro:Primary/Unknown ds:UpToDate/DUnknown   r-----
    ns:0 nr:0 dw:600 dr:263205 al:0 bm:0 lo:0 pe:0 ua:0 ap:0 ep:1 wo:f oos:696
 1: cs:SyncTarget ro:Primary/Primary ds:Inconsistent/UpToDate C r-----
    ns:0 nr:5327932 dw:5327616 dr:32 al:0 bm:325 lo:2 pe:7494 ua:2 ap:1 ep:1
wo:f oos:853695940
        [>....................] sync'ed:  0.7% (833684/838888)Mfinish: 2:03:44
speed: 114,984 (108,724) want: 1,000,001 K/sec

10: cs:Connected ro:Primary/Primary ds:UpToDate/UpToDate C r-----
    ns:84 nr:12 dw:44 dr:263073 al:0 bm:3 lo:0 pe:0 ua:0 ap:0 ep:1 wo:f oos:0
root@cluster-data-1:~# drbd-overview
  0:system_data  StandAlone Primary/Unknown UpToDate/DUnknown     r-----
/mnt/system ocfs2     90G   862M 90G 1%
  1:vm_volumes   SyncTarget Primary/Primary Inconsistent/UpToDate C      r-----

        [>....................] sync'ed:  0.7% (833252/838888)Mfinish: 1:58:12
speed: 120,292 (108,924) want: 1,000,001 K/sec
 10:repository   Connected  Primary/Primary UpToDate/UpToDate     C      r-----
     /mnt/repo ocfs2 250G 82G 169G 33%
root@cluster-data-1:~# drbd-overview
  0:system_data  StandAlone Primary/Unknown UpToDate/DUnknown     r-----
/mnt/system ocfs2     90G   862M 90G 1%
  1:vm_volumes   SyncTarget Primary/Primary Inconsistent/UpToDate C      r-----

        [>....................] sync'ed:  1.0% (830752/838888)Mfinish: 2:01:35
speed: 116,584 (109,628) want: 1,000,001 K/sec
 10:repository   Connected  Primary/Primary UpToDate/UpToDate     C      r-----
     /mnt/repo ocfs2 250G 82G 169G 33%
root@cluster-data-1:~# crm resource restart ms_drbd_system_data
INFO: ordering ms_drbd_system_data to stop
waiting for stop to finish
......................................................................................................................................................................................................................................................................................................................................................................................
done
INFO: ordering ms_drbd_system_data to start
root@cluster-data-1:~# drbd-overview
  0:system_data  StandAlone Primary/Unknown UpToDate/DUnknown     r-----
/mnt/system ocfs2     90G   862M 90G 1%
  1:vm_volumes   SyncTarget Primary/Primary Inconsistent/UpToDate C      r-----

        [>....................] sync'ed:  4.3% (803584/838888)Mfinish: 1:58:20
speed: 115,884 (112,964) want: 1,000,001 K/sec
 10:repository   Connected  Primary/Primary UpToDate/UpToDate     C      r-----
     /mnt/repo ocfs2 250G 82G 169G 33%
root@cluster-data-1:~# crm configure show ms_drbd_system_data
ms ms_drbd_system_data p_drbd_system_data \
        meta notify="true" clone-max="2" master-max="2" target-role="Started"
root@cluster-data-1:~# crm configure edit ms_drbd_system_data
root@cluster-data-1:~# crm configure show ms_drbd_system_data
ms ms_drbd_system_data p_drbd_system_data \
        meta notify="true" clone-max="2" master-max="1" target-role="Started"
root@cluster-data-1:~# crm resource restart ms_drbd_system_data
INFO: ordering ms_drbd_system_data to stop
waiting for stop to finish
.................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................^CCtrl-C,
leaving
root@cluster-data-1:~# drbd-overview
  0:system_data  StandAlone Primary/Unknown UpToDate/DUnknown     r-----
/mnt/system ocfs2     90G   862M 90G 1%
  1:vm_volumes   SyncTarget Primary/Primary Inconsistent/UpToDate C      r-----

        [=>..................] sync'ed: 10.2% (754088/838888)Mfinish: 1:52:40
speed: 114,200 (113,804) want: 1,000,001 K/sec
 10:repository   Connected  Primary/Primary UpToDate/UpToDate     C      r-----
     /mnt/repo ocfs2 250G 82G 169G 33%
root@cluster-data-1:~# crm resource stop ms_drbd_system_data
root@cluster-data-1:~# drbd-overview
  0:system_data  StandAlone Primary/Unknown UpToDate/DUnknown     r-----
/mnt/system ocfs2     90G   862M 90G 1%
  1:vm_volumes   SyncTarget Primary/Primary Inconsistent/UpToDate C      r-----

        [=>..................] sync'ed: 10.5% (750956/838888)Mfinish: 1:51:34
speed: 114,840 (113,688) want: 1,000,001 K/sec
 10:repository   Connected  Primary/Primary UpToDate/UpToDate     C      r-----
     /mnt/repo ocfs2 250G 82G 169G 33%
root@cluster-data-1:~# crm resource stop ms_drbd_system_data
root@cluster-data-1:~# drbd-overview
  0:system_data  StandAlone Primary/Unknown UpToDate/DUnknown     r-----
/mnt/system ocfs2     90G   862M 90G 1%
  1:vm_volumes   SyncTarget Primary/Primary Inconsistent/UpToDate C      r-----

        [=>..................] sync'ed: 10.9% (748088/838888)Mfinish: 1:46:49
speed: 119,500 (113,804) want: 1,000,001 K/sec
 10:repository   Connected  Primary/Primary UpToDate/UpToDate     C      r-----
     /mnt/repo ocfs2 250G 82G 169G 33%



Настойки выглядят адекватно, но смущает одна вещь - у вас вообще выключены все хендлеры - так и задумано?

На счет переодическое выпадения в Unknown - оно так с обоих нод выглядит? Что в dmesg в районе split-brain?

У меня две версии - проблема с сетью (что маловероятно) или диски не успевают (вполне вероятно)

Reply to: