[Gluster-users] The dht-layout interval is missing

jifeng-call 17607319886 at 163.com
Fri May 29 07:56:18 UTC 2020


Hi All,




I have 6 servers that form a glusterfs 2x3 distributed replication volume, the details are as follows:




[root at node1 ~]# gluster volume info

Volume Name: abcd_vol

Type: Distributed-Replicate

Volume ID: c9848daa-b06f-4f82-a2f8-1b425b8e869c

Status: Started

Snapshot Count: 0

Number of Bricks: 2 x 3 = 6

Transport-type: tcp

Bricks:

Brick1: node1:/export/test/abcd

Brick2: node2:/export/test/abcd

Brick3: node3:/export/test/abcd

Brick4: node4:/export/test/abcd

Brick5: node5:/export/test/abcd

Brick6: node6:/export/test/abcd

Options Reconfigured:

diagnostics.client-log-level: DEBUG

diagnostics.brick-log-level: DEBUG

performance.client-io-threads: off

nfs.disable: on

transport.address-family: inet

cluster.quorum-type: auto




Restart 6 servers at a time, the server automatically mounts the abcd_vol volume to the / home / test-abcd directory during the startup process:




[root at node1 ~]# df -h | grep abcd_vol

node1:/abcd_vol            4.5T  362G  4.1T    9% /home/test-abcd




Failed to create a file in the glusterfs mount directory of node3, the log shows the error as follows (the files created by other servers are normal):




[2020-05-29 05:55:03.065656] E [MSGID: 109011] [dht-common.c:8683:dht_create] 0-abcd_vol-dht: no subvolume in layout for path=/.abcd.time.405fa8a6-575d-4474-a97f-a107cbf1c673.18512

[2020-05-29 05:55:03.065719] W [fuse-bridge.c:2122:fuse_create_cbk] 0-glusterfs-fuse: 2454790: /.abcd.time.405fa8a6-575d-4474-a97f-a107cbf1c673.18512 => -1 (输入/输出错误)

[2020-05-29 05:55:03.680303] W [MSGID: 109011] [dht-layout.c:186:dht_layout_search] 0-abcd_vol-dht: no subvolume for hash (value) = 1738969696

[2020-05-29 05:55:04.687456] W [MSGID: 109011] [dht-layout.c:186:dht_layout_search] 0-abcd_vol-dht: no subvolume for hash (value) = 491618322

[2020-05-29 05:55:04.688612] W [MSGID: 109011] [dht-layout.c:186:dht_layout_search] 0-abcd_vol-dht: no subvolume for hash (value) = 1738969696

[2020-05-29 05:55:05.694446] W [MSGID: 109011] [dht-layout.c:186:dht_layout_search] 0-abcd_vol-dht: no subvolume for hash (value) = 1738969696

[2020-05-29 05:55:05.830555] W [MSGID: 109011] [dht-layout.c:186:dht_layout_search] 0-abcd_vol-dht: no subvolume for hash (value) = 491618322

[2020-05-29 05:55:06.700423] W [MSGID: 109011] [dht-layout.c:186:dht_layout_search] 0-abcd_vol-dht: no subvolume for hash (value) = 1738969696

[2020-05-29 05:55:07.706536] W [MSGID: 109011] [dht-layout.c:186:dht_layout_search] 0-abcd_vol-dht: no subvolume for hash (value) = 1738969696

[2020-05-29 05:55:07.833049] W [MSGID: 109011] [dht-layout.c:186:dht_layout_search] 0-abcd_vol-dht: no subvolume for hash (value) = 491618322

[2020-05-29 05:55:08.712128] W [MSGID: 109011] [dht-layout.c:186:dht_layout_search] 0-abcd_vol-dht: no subvolume for hash (value) = 1738969696

The message "W [MSGID: 109011] [dht-layout.c:186:dht_layout_search] 0-abcd_vol-dht: no subvolume for hash (value) = 1738969696" repeated 2 times between [2020-05-29 05:55:08.712128] and [2020-05-29 05:55:09.718541]




The pathinfo information in the / home / test-abcd directory is displayed as follows:




[root at node3 test-abcd]# getfattr -d -m .  -n trusted.glusterfs.pathinfo /home/test-abcd/

getfattr: Removing leading '/' from absolute path names

# file: home/test-abcd/

trusted.glusterfs.pathinfo="((<DISTRIBUTE:abcd_vol-dht> (<REPLICATE:abcd_vol-replicate-0> <POSIX(/export/test/abcd):node2:/export/test/abcd/> <POSIX(/export/test/abcd):node3:/export/test/abcd/> <POSIX(/export/test/abcd):node1:/export/te

st/abcd/>) (<REPLICATE:abcd_vol-replicate-1> <POSIX(/export/test/abcd):node5:/export/test/abcd/> <POSIX(/export/test/abcd):node6:/export/test/abcd/> <POSIX(/export/test/abcd):node4:/export/test/abcd/>)) (abcd_vol-dht-layout (abcd_vol-re

plicate-0 0 0) (abcd_vol-replicate-1 3539976838 4294967295)))"




It can be seen from the above information that abcd_vol-dht-layout is missing the interval 0 to 3539976837




After umount, remounting returned to normal ... What is the reason?




Best regards
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gluster.org/pipermail/gluster-users/attachments/20200529/b6d1768b/attachment.html>


More information about the Gluster-users mailing list