[Gluster-users] postgresql update error when glusterfs as the postgresql data directory

[hao.wu]

hi guys
I have 4 servers with a RAID0 (4 * 600GB sas hdds) each.  OS is centos 6.6, kerel version is 2.6.32-504.el6.x86_64. The disk in every server is partitioned as the below:
/dev/sda1 /boot
/dev/sda2 lvm(/ and swap)
/dev/sda3 ext4(this is for the glusterfs brick)
then I made a distributed volume gv2, the info is the below:
root at pc-zjdats14 ~]# gluster volume info gv2
 
Volume Name: gv2
Type: Distribute
Volume ID: 8587e54c-2179-49f5-b1b7-8a6b6c265ab1
Status: Started
Number of Bricks: 4
Transport-type: tcp
Bricks:
Brick1: pc-zjdats14:/data/brick1/gv2
Brick2: pc-zjdats15:/data/brick1/gv2
Brick3: pc-zjdats16:/data/brick1/gv2
Brick4: pc-zjdats17:/data/brick1/gv2
then I mounted the gv2 to /shrfs mount point in node1(pc-zjdats14). only node1 mounted the glusterfs.
I installed the postgresql in the / filesystem and data in the /shrfs glusterfs.
then sometimes I update the tables in the postgresql,   the error will occur. and some error in the glusterd log like the below
[2015-10-15 01:06:12.542793] E [iobuf.c:759:iobuf_unref] (--> /usr/lib64/libglusterfs.so.0(_gf_log_callingfn+0x1e0)[0x7f52649d3580] (--> /usr/lib64/glusterfs/3.6.6/xlator/performance/quick-read.so(qr_readv_cached+0xb7)[0x7f5253336ea7] (--> /usr/lib64/glusterfs/3.6.6/xlator/performance/quick-read.so(qr_readv+0x62)[0x7f52533374c2] (--> /usr/lib64/libglusterfs.so.0(default_readv_resume+0x14d)[0x7f52649dd75d] (--> /usr/lib64/libglusterfs.so.0(call_resume+0x3d6)[0x7f52649f54b6] ))))) 0-iobuf: invalid argument: iobuf
[2015-10-15 01:06:12.542988] E [iobuf.c:865:iobref_unref] (--> /usr/lib64/libglusterfs.so.0(_gf_log_callingfn+0x1e0)[0x7f52649d3580] (--> /usr/lib64/glusterfs/3.6.6/xlator/performance/quick-read.so(qr_readv_cached+0xc1)[0x7f5253336eb1] (--> /usr/lib64/glusterfs/3.6.6/xlator/performance/quick-read.so(qr_readv+0x62)[0x7f52533374c2] (--> /usr/lib64/libglusterfs.so.0(default_readv_resume+0x14d)[0x7f52649dd75d] (--> /usr/lib64/libglusterfs.so.0(call_resume+0x3d6)[0x7f52649f54b6] ))))) 0-iobuf: invalid argument: iobref
[2015-10-15 01:06:12.543303] E [iobuf.c:759:iobuf_unref] (--> /usr/lib64/libglusterfs.so.0(_gf_log_callingfn+0x1e0)[0x7f52649d3580] (--> /usr/lib64/glusterfs/3.6.6/xlator/performance/quick-read.so(qr_readv_cached+0xb7)[0x7f5253336ea7] (--> /usr/lib64/glusterfs/3.6.6/xlator/performance/quick-read.so(qr_readv+0x62)[0x7f52533374c2] (--> /usr/lib64/libglusterfs.so.0(default_readv_resume+0x14d)[0x7f52649dd75d] (--> /usr/lib64/libglusterfs.so.0(call_resume+0x3d6)[0x7f52649f54b6] ))))) 0-iobuf: invalid argument: iobuf


So,  what is the glusterfs error and what cause it? how do I resolve the error ?
any response will be appreciated. Thanks.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.gluster.org/pipermail/gluster-users/attachments/20151023/76793c3a/attachment.html>