PV status unavailable

[정락교]

Error Message



===============================================================

Feb 20 17:09:22 vmunix: SCSI: Async write error -- dev: b 31 0x053000, errno: 126, resid: 4096,

Feb 20 17:09:22 vmunix: blkno: 19412820, sectno: 38825640, offset: 2698858496, bcount: 4096.

Feb 20 17:09:22 vmunix: SCSI: Async write error -- dev: b 31 0x053000, errno: 126, resid: 57344,

Feb 20 17:09:22 vmunix: blkno: 21725248, sectno: 43450496, offset: 771817472, bcount: 57344.

Feb 20 17:09:22 vmunix: SCSI: Async write error -- dev: b 31 0x053000, errno: 126, resid: 8192,

Feb 20 17:09:22 vmunix: blkno: 21725304, sectno: 43450608, offset: 771874816, bcount: 8192.

Feb 20 17:09:22 vmunix: SCSI: Read error -- dev: b 31 0x053000, errno: 126, resid: 1024,

Feb 20 17:09:22 vmunix: blkno: 8, sectno: 16, offset: 8192, bcount: 1024.

Feb 20 17:09:22 vmunix: LVM: VG 64 0x050000: Lost quorum.

Feb 20 17:09:22 vmunix:

Feb 20 17:09:22 above message repeats 3 times

Feb 20 17:09:22 vmunix: This may block configuration changes and I/Os. In order to reestablish quorum at least 1 of the following PVs (represented by current link) must become available:

Feb 20 17:09:22 vmunix: <31 0x053000>

Feb 20 17:09:22 vmunix: LVM: VG 64 0x050000: PVLink 31 0x053000 Failed! The PV is not accessible.

Feb 20 17:09:27 vmunix: LVM: VG 64 0x050000: Reestablished quorum.

Feb 20 17:09:27 vmunix: LVM: VG 64 0x050000: PVLink 31 0x053000 Recovered.

Feb 20 17:10:31 vmunix:

Feb 20 17:10:31 vmunix: SCSI: Async write error -- dev: b 31 0x053000, errno: 126, resid: 57344,

Feb 20 17:10:31 vmunix: blkno: 21594944, sectno: 43189888, offset: 638386176, bcount: 57344.

Feb 20 17:11:35 vmunix: SCSI: Async write error -- dev: b 31 0x053000, errno: 126, resid: 4096,

Feb 20 17:11:35 vmunix: blkno: 17265984, sectno: 34531968, offset: 500498432, bcount: 4096.

Feb 20 17:11:35 vmunix: blkno: 26540, sectno: 53080, offset: 27176960, bcount: 4096.

Feb 20 17:11:35 vmunix: SCSI: Async write error -- dev: b 31 0x053000, errno: 126, resid: 32768,

Feb 20 17:11:35 vmunix: SCSI: Async write error -- dev: b 31 0x053000, errno: 126, resid: 4096,

Feb 20 17:11:35 vmunix: blkno: 21430912, sectno: 42861824, offset: 470417408, bcount: 32768.

Feb 20 17:11:35 vmunix: SCSI: Async write error -- dev: b 31 0x053000, errno: 126, resid: 24576,

Feb 20 17:11:35 vmunix: blkno: 21430944, sectno: 42861888, offset: 470450176, bcount: 24576.

Feb 20 17:12:39 vmunix: SCSI: Async write error -- dev: b 31 0x053000, errno: 126, resid: 40960,

Feb 20 17:12:39 vmunix: blkno: 21395792, sectno: 42791584, offset: 434454528, bcount: 40960.

Feb 20 17:12:39 vmunix:

Feb 20 17:13:34 above message repeats 5 times

Feb 20 17:13:43 vmunix:

Feb 20 17:13:43 vmunix: SCSI: Async write error -- dev: b 31 0x053000, errno: 126, resid: 32768,

Feb 20 17:13:43 vmunix: blkno: 20514784, sectno: 41029568, offset: 3827269632, bcount: 32768.

Feb 20 17:13:43 vmunix: SCSI: Async write error -- dev: b 31 0x053000, errno: 126, resid: 65536,

Feb 20 17:13:43 vmunix: blkno: 19974464, sectno: 39948928, offset: 3273981952, bcount: 65536.

Feb 20 17:13:43 vmunix: SCSI: Read error -- dev: b 31 0x053000, errno: 126, resid: 1024,

Feb 20 17:13:43 vmunix: blkno: 8, sectno: 16, offset: 8192, bcount: 1024.

Feb 20 17:13:43 vmunix: LVM: VG 64 0x050000: Lost quorum.

Feb 20 17:13:43 vmunix:

Feb 20 17:13:43 above message repeats 2 times

Feb 20 17:13:43 vmunix: This may block configuration changes and I/Os. In order to reestablish quorum at least 1 of the following PVs (represented by current link) must become available:

Feb 20 17:13:43 vmunix: <31 0x053000>

Feb 20 17:13:43 vmunix: LVM: VG 64 0x050000: PVLink 31 0x053000 Failed! The PV is not accessible.

Feb 20 17:13:48 vmunix: LVM: VG 64 0x050000: Reestablished quorum.

Feb 20 17:13:48 vmunix: LVM: VG 64 0x050000: PVLink 31 0x053000 Recovered.

Feb 20 17:14:52 vmunix:

Feb 20 17:14:52 vmunix: SCSI: Read error -- dev: b 31 0x053000, errno: 126, resid: 1024,

Feb 20 17:14:52 vmunix: blkno: 8, sectno: 16, offset: 8192, bcount: 1024.

Feb 20 17:14:52 vmunix: LVM: VG 64 0x050000: Lost quorum.

Feb 20 17:14:52 vmunix: This may block configuration changes and I/Os. In order to reestablish quorum at least 1 of the following PVs (represented by current link) must become available:

Feb 20 17:14:52 vmunix: <31 0x053000>

Feb 20 17:14:52 vmunix: LVM: VG 64 0x050000: PVLink 31 0x053000 Failed! The PV is not accessible.

Feb 20 17:14:57 vmunix: LVM: VG 64 0x050000: Reestablished quorum.

Feb 20 17:14:57 vmunix: LVM: VG 64 0x050000: PVLink 31 0x053000 Recovered.

Feb 20 17:16:01 vmunix:

Feb 20 17:16:01 vmunix: SCSI: Async write error -- dev: b 31 0x053000, errno: 126, resid: 57344,

Feb 20 17:16:01 vmunix: blkno: 21725248, sectno: 43450496, offset: 771817472, bcount: 57344.

Feb 20 17:16:01 vmunix: SCSI: Async write error -- dev: b 31 0x053000, errno: 126, resid: 8192,

Feb 20 17:16:01 vmunix: blkno: 21725304, sectno: 43450608, offset: 771874816, bcount: 8192.

Feb 20 17:16:01 vmunix:

Feb 20 17:17:05 vmunix:

Feb 20 17:17:05 vmunix: SCSI: Read error -- dev: b 31 0x053000, errno: 126, resid: 1024,

Feb 20 17:17:05 vmunix: blkno: 8, sectno: 16, offset: 8192, bcount: 1024.

Feb 20 17:17:05 vmunix: LVM: VG 64 0x050000: Lost quorum.

Feb 20 17:17:05 vmunix: This may block configuration changes and I/Os. In order to reestablish quorum at least 1 of the following PVs (represented by current link) must become available:

Feb 20 17:17:05 vmunix: <31 0x053000>

Feb 20 17:17:05 vmunix: LVM: VG 64 0x050000: PVLink 31 0x053000 Failed! The PV is not accessible.

Feb 20 17:17:10 vmunix: LVM: VG 64 0x050000: Reestablished quorum.

Feb 20 17:17:10 vmunix: LVM: VG 64 0x050000: PVLink 31 0x053000 Recovered.

Feb 20 17:18:14 vmunix:

Feb 20 17:18:14 vmunix: SCSI: Read error -- dev: b 31 0x053000, errno: 126, resid: 1024,

Feb 20 17:18:14 vmunix: blkno: 8, sectno: 16, offset: 8192, bcount: 1024.

Feb 20 17:18:14 vmunix: LVM: VG 64 0x050000: Lost quorum.

Feb 20 17:18:14 vmunix: This may block configuration changes and I/Os. In order to reestablish quorum at least 1 of the following PVs (represented by current link) must become available:

Feb 20 17:18:14 vmunix: <31 0x053000>

Feb 20 17:18:14 vmunix: LVM: VG 64 0x050000: PVLink 31 0x053000 Failed! The PV is not accessible.

Feb 20 17:18:19 vmunix: LVM: VG 64 0x050000: Reestablished quorum.

Feb 20 17:18:19 vmunix: LVM: VG 64 0x050000: PVLink 31 0x053000 Recovered.

Feb 20 17:19:23 vmunix:

Feb 20 17:19:23 vmunix: SCSI: Async write error -- dev: b 31 0x053000, errno: 126, resid: 57344,

Feb 20 17:19:23 vmunix: blkno: 21594944, sectno: 43189888, offset: 638386176, bcount: 57344.

Feb 20 17:20:27 vmunix: SCSI: Read error -- dev: b 31 0x053000, errno: 126, resid: 1024,

Feb 20 17:20:27 vmunix: blkno: 8, sectno: 16, offset: 8192, bcount: 1024.

Feb 20 17:20:27 vmunix: LVM: VG 64 0x050000: Lost quorum.

Feb 20 17:20:27 vmunix:

Feb 20 17:20:27 vmunix: This may block configuration changes and I/Os. In order to reestablish quorum at least 1 of the following PVs (represented by current link) must become available:

Feb 20 17:20:27 vmunix: <31 0x053000>

Feb 20 17:20:27 vmunix: LVM: VG 64 0x050000: PVLink 31 0x053000 Failed! The PV is not accessible.

Feb 20 17:20:32 vmunix: LVM: VG 64 0x050000: Reestablished quorum.

Feb 20 17:20:32 vmunix: LVM: VG 64 0x050000: PVLink 31 0x053000 Recovered.

Feb 20 17:21:36 vmunix:

Feb 20 17:21:36 vmunix: SCSI: Read error -- dev: b 31 0x053000, errno: 126, resid: 1024,

Feb 20 17:21:36 vmunix: blkno: 8, sectno: 16, offset: 8192, bcount: 1024.

Feb 20 17:21:36 vmunix: LVM: VG 64 0x050000: Lost quorum.

Feb 20 17:21:36 vmunix: This may block configuration changes and I/Os. In order to reestablish quorum at least 1 of the following PVs (represented by current link) must become available:

Feb 20 17:21:36 vmunix: <31 0x053000>

Feb 20 17:21:36 vmunix: LVM: VG 64 0x050000: PVLink 31 0x053000 Failed! The PV is not accessible.

Feb 20 17:21:41 vmunix: LVM: VG 64 0x050000: Reestablished quorum.

Feb 20 17:21:41 vmunix: LVM: VG 64 0x050000: PVLink 31 0x053000 Recovered.

Feb 20 17:22:45 vmunix:

Feb 20 17:22:45 vmunix: SCSI: Read error -- dev: b 31 0x053000, errno: 126, resid: 1024,

Feb 20 17:22:45 vmunix: blkno: 8, sectno: 16, offset: 8192, bcount: 1024.

Feb 20 17:22:45 vmunix: LVM: VG 64 0x050000: Lost quorum.

Feb 20 17:22:45 vmunix: This may block configuration changes and I/Os. In order to reestablish quorum at least 1 of the following PVs (represented by current link) must become available:

Feb 20 17:22:45 vmunix: <31 0x053000>

Feb 20 17:22:45 vmunix: LVM: VG 64 0x050000: PVLink 31 0x053000 Failed! The PV is not accessible.

Feb 20 17:22:50 vmunix: LVM: VG 64 0x050000: Reestablished quorum.

Feb 20 17:22:50 vmunix: LVM: VG 64 0x050000: PVLink 31 0x053000 Recovered.

Feb 20 17:23:54 vmunix:

Feb 20 17:23:54 vmunix: SCSI: Read error -- dev: b 31 0x053000, errno: 126, resid: 1024,

Feb 20 17:23:54 vmunix: blkno: 8, sectno: 16, offset: 8192, bcount: 1024.

Feb 20 17:23:54 vmunix: LVM: VG 64 0x050000: Lost quorum.

Feb 20 17:23:54 vmunix: This may block configuration changes and I/Os. In order to reestablish quorum at least 1 of the following PVs (represented by current link) must become available:

Feb 20 17:23:54 vmunix: <31 0x053000>

Feb 20 17:23:54 vmunix: LVM: VG 64 0x050000: PVLink 31 0x053000 Failed! The PV is not accessible.

Feb 20 17:23:59 vmunix: LVM: VG 64 0x050000: Reestablished quorum.

Feb 20 17:23:59 vmunix: LVM: VG 64 0x050000: PVLink 31 0x053000 Recovered.

Feb 20 17:25:03 vmunix:

Feb 20 17:25:03 vmunix: SCSI: Read error -- dev: b 31 0x053000, errno: 126, resid: 1024,

Feb 20 17:25:03 vmunix: blkno: 8, sectno: 16, offset: 8192, bcount: 1024.

Feb 20 17:25:03 vmunix: LVM: VG 64 0x050000: Lost quorum.

Feb 20 17:25:03 vmunix: This may block configuration changes and I/Os. In order to reestablish quorum at least 1 of the following PVs (represented by current link) must become available:

Feb 20 17:25:03 vmunix: <31 0x053000>

Feb 20 17:25:03 vmunix: LVM: VG 64 0x050000: PVLink 31 0x053000 Failed! The PV is not accessible.

Feb 20 17:25:08 vmunix: LVM: VG 64 0x050000: Reestablished quorum.

Feb 20 17:25:08 vmunix: LVM: VG 64 0x050000: PVLink 31 0x053000 Recovered.

Feb 20 17:26:12 vmunix:

Feb 20 17:26:12 vmunix: SCSI: Read error -- dev: b 31 0x053000, errno: 126, resid: 1024,

Feb 20 17:26:12 vmunix: blkno: 8, sectno: 16, offset: 8192, bcount: 1024.

Feb 20 17:26:12 vmunix: LVM: VG 64 0x050000: Lost quorum.

Feb 20 17:26:12 vmunix: This may block configuration changes and I/Os. In order to reestablish quorum at least 1 of the following PVs (represented by current link) must become available:

Feb 20 17:26:12 vmunix: <31 0x053000>

Feb 20 17:26:12 vmunix: LVM: VG 64 0x050000: PVLink 31 0x053000 Failed! The PV is not accessible.



==================================================================



고수님들의 도움 부탁드립니다.

외장스토리지 볼륨을 연결해서 쓰고 있습니다.

위와 같이 vg05 볼륨이 SCSI Error 가 나옵니다.

vgdisplay -v vg05 하면 vg status는 available인데, pv status가 unavailable가 되면서 먹통이 되어 버립니다.

평상시엔 나오지 않다가 큰 파일을 백업을 한다던가 하는 작업을 할 때면 위와 같은 메시지가 뜨면서 먹통이 되어 버립니다.

물론 대용량 파일로 파일시스템을 만들었구요.

무엇이 문제인가요???



답변 좀 부탁드립니다.

[양계전]

외장 디스크 터미네이터 타입을 확인 하셔야 할듯합니다



예를 들어



LVD_SE 인지 그냥 SE인지



LVD_SE Type 외장 디스크에 SE 타입을 끼우면 SE만큼 속도가 줄며



요청 속도보다 나오지 않을때 위와같은 메세지가 나올경우도 있습니다..



일 예 이므로 참고만 하세요

[정락교]

외장스토리지는 Qlogic HBA, FC type 으로 연결되어 있습니다.

[양계전]

그럼 혹시 Tape 장치와 같이 묶여있는지는 않는지요 HBA 카드에요

[정락교]

Tape 장치는 외장형 20/40 DAT장치로 SCSI로 연결되어 있구요, 외장스토리지는 Fiber Type의 HBA로 연결되어 있습니다.

[고재진]

안녕하세요.



스토리지가 이중화가 되어있다는것을 감안하여 pvlink가 되어있다면

위 로그는 한쪽 path에서만 이런 문제가 생기는 것입니다.

HBA 카드나 fc 케이블등 H/W 쪽 문제로 생각 할 수있습니다.



PVlink 가 되어있지 않고

대용량의 백업을 받을 때만 로그가 생기면

I/O traffic 이 과중해서 병목현상이 생기는건 아닌지 생각해봐야됩니다. 백업을 받을때 i/o performance를 확인해보고 과중하다 싶으면 분산시키는 방법을 생각해야 될듯하네요.



그리고 문제가 명확하지 않은경우

우선 패치를 올려놓는것도 좋은 방법입니다.



패치를 최신으로 올리시든지



fc 패치

lvm 패치

scsi 패치



를 고려하셔야 될것 같습니다.



그럼.

[정락교]

고재진님!

아래의 패치는 어디서 찾아야 하고, 어떤걸 받아야 하나요?

서버는 rp5405 Model이고, v11.11 버전입니다.



fc 패치

lvm 패치

scsi 패치

[이태곤]

안녕하십니까



평상시에는 정상이면 부하분산을 생각해보셔야 합니다.



예전에 EVA4000에 pclink만으로 구성되어 있었는데 백업만 받으면 위와 같은 에러가 발생했는데 autopath을 설치고 문제가 발생하지 않았습니다.



스토리지가 어떤것인지는 모르겠지만 hp의 autopath, EMC의 powerpath을 권고합니다.

만약 Pvlink만으로 구성되어 있다면 lvol을 만드실때 스트라이핑을 걸어주시면 됩니다.



현재상황에서 가장먼저 해봐야 할것은 pvchange로 time out값을 늘려보십시오



# pvchange -t 180 /dev/dsk/5t3d0



이후 에러메시지가 발생된다면 부하분산을 하셔야 합니다.



스토리지에 큰 볼륨을 하나 만드는것보다 두개을 만들고 vgextend을 하시고 다시 lvcreate하실때 스트라이핑을 거시면 HBA두개을 모두 사용할수 있을것입니다.



가장쉬운 방법은 autopath, powerpath일것 같습니다



# lvdisplay -v /dev/vg01/lvol1



--- Distribution of logical volume ---

PV Name LE on PV PE on PV

/dev/dsk/c5t1d0 32 32

/dev/dsk/c6t0d1 32 32



--- Logical extents ---

LE PV1 PE1 Status 1

00000 /dev/dsk/c5t1d0 00000 current

00001 /dev/dsk/c6t0d1 00001 current

00002 /dev/dsk/c5t1d0 00002 current

00003 /dev/dsk/c6t0d1 00003 current

00004 /dev/dsk/c5t1d0 00004 current

00005 /dev/dsk/c6t0d1 00005 current





그리고 FC, LVM, SCSI는 itrc.hp.com에서 각각의 단어로 검색하시면 됨니다

Fibre Channel Mass Storage Patch 등등