|
|
Sponsored Content
Operating Systems
Solaris
puzzled with VxVM and iostat..
Post 302241055 by samar on Sunday 28th of September 2008 04:31:05 AM
09-28-2008
puzzled with VxVM and iostat..
Hi all,
One disk on my root disk group failed in Veritas Volume manager.
I replaced it with new one, initialized it and placed it with removed one.
it Synchronized plexes and everything is fine. this node was second standby node of Sun cluster. yesterday I had failure on active node with boot disk (same as second node) .. but what is puzzled me here it has mirror and it was alive .. it gives me warning not even error like:
Sep 27 22:07:19 host-1 scsi: [ID 107833 kern.warning] WARNING: /ssm@0,0/pci@18,700000/scsi@2/sd@0,0 (sd3):
Sep 27 22:07:19 host-1 disk not responding to selection
Sep 27 22:07:19 host-1 vxio: [ID 107055 kern.warning] WARNING: VxVM vxio V-5-3-0 voldmp_errbuf_sio_start: Failed to flush the error buffer 0xdbf3c330 on device0x12800000003 to DMP
Sep 27 22:07:19 host-1 vxio: [ID 771159 kern.warning] WARNING: VxVM vxio V-5-0-2 Subdisk rootdg_2-01 block 64703344: Uncorrectable read error
First 2 lines are from system not from VxVM but after that vxio informed.
So it was at 22:07:19 and 30 seconds it continued give this log to terminal after what cluster said :
Sep 27 22:07:59 host-1 cl_dlpitrans: [ID 624622 kern.notice] Notifying cluster that this node is panicking
Sep 27 22:07:59 host-1 unix: [ID 836849 kern.notice]
Sep 27 22:07:59 host-1 ^Mpanic[cpu0]/thread=2a100045d20:
Sep 27 22:07:59 host-1 unix: [ID 562397 kern.notice] Failfast: Aborting because "rgmd" died 30 seconds ago.
and went to reboot ..
So why cluster got this warning like whole root-dg failured and went to reboot?!!!
and after that Application swithed to second node (where I changed failured disk not long ago and everything is fine) .. so there it started normally but I see in iostat that on first disk (exactly that was replaced) ..it has very big IO than other mirrored disk. So how Veritas synchronize, I mean how it works. Mirrored disks in VxVM always must have same IO ???
it is some time :
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
0.4 1.7 17.1 28.0 0.2 0.0 90.7 14.3 19 1 c1t0d0
1.1 2.1 61.3 17.5 0.0 0.0 0.0 11.8 0 1 c1t1d0
and sometimes:
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
0.0 0.2 0.0 1.6 1.0 0.0 5000.0 5.5 100 0 c1t0d0
0.0 0.2 0.0 1.6 0.0 0.0 0.0 5.4 0 0 c1t1d0
??!!!!!!!!
Thanks all
One disk on my root disk group failed in Veritas Volume manager.
I replaced it with new one, initialized it and placed it with removed one.
it Synchronized plexes and everything is fine. this node was second standby node of Sun cluster. yesterday I had failure on active node with boot disk (same as second node) .. but what is puzzled me here it has mirror and it was alive .. it gives me warning not even error like:
Sep 27 22:07:19 host-1 scsi: [ID 107833 kern.warning] WARNING: /ssm@0,0/pci@18,700000/scsi@2/sd@0,0 (sd3):
Sep 27 22:07:19 host-1 disk not responding to selection
Sep 27 22:07:19 host-1 vxio: [ID 107055 kern.warning] WARNING: VxVM vxio V-5-3-0 voldmp_errbuf_sio_start: Failed to flush the error buffer 0xdbf3c330 on device0x12800000003 to DMP
Sep 27 22:07:19 host-1 vxio: [ID 771159 kern.warning] WARNING: VxVM vxio V-5-0-2 Subdisk rootdg_2-01 block 64703344: Uncorrectable read error
First 2 lines are from system not from VxVM but after that vxio informed.
So it was at 22:07:19 and 30 seconds it continued give this log to terminal after what cluster said :
Sep 27 22:07:59 host-1 cl_dlpitrans: [ID 624622 kern.notice] Notifying cluster that this node is panicking
Sep 27 22:07:59 host-1 unix: [ID 836849 kern.notice]
Sep 27 22:07:59 host-1 ^Mpanic[cpu0]/thread=2a100045d20:
Sep 27 22:07:59 host-1 unix: [ID 562397 kern.notice] Failfast: Aborting because "rgmd" died 30 seconds ago.
and went to reboot ..
So why cluster got this warning like whole root-dg failured and went to reboot?!!!
and after that Application swithed to second node (where I changed failured disk not long ago and everything is fine) .. so there it started normally but I see in iostat that on first disk (exactly that was replaced) ..it has very big IO than other mirrored disk. So how Veritas synchronize, I mean how it works. Mirrored disks in VxVM always must have same IO ???
it is some time :
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
0.4 1.7 17.1 28.0 0.2 0.0 90.7 14.3 19 1 c1t0d0
1.1 2.1 61.3 17.5 0.0 0.0 0.0 11.8 0 1 c1t1d0
and sometimes:
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
0.0 0.2 0.0 1.6 1.0 0.0 5000.0 5.5 100 0 c1t0d0
0.0 0.2 0.0 1.6 0.0 0.0 0.0 5.4 0 0 c1t1d0
??!!!!!!!!
Thanks all
|
|
10 More Discussions You Might Find Interesting
1. Programming
I hava been reading AUPE these days. I really am puzzled with the presentation of real user(group) ID, effective user(group) ID. How do they effect on the execution of process? What's the relationship between them? Appreciate your help. (4 Replies)
Discussion started by: lethefe
4 Replies
2. Programming
#include <iostream.h>
class A
{
public:
void f(void)
{
cout << "hello world \n" ;
}
};
void main()
{
A *a;
a = 0 ;
a->f(); // OOPs...Am I mad? What am I going to do ?
} (1 Reply)
Discussion started by: RipClaw
1 Replies
3. Filesystems, Disks and Memory
All,
I am attempting to help tune a Sun for better performance (mainly for SAS 9.1), and have found indicators pointing to poor I/O utilization. I have run iostat -cx, and found one device in particular where the %w is in the 90's during processing. I have a feeling that this is where the SAS... (3 Replies)
Discussion started by: dj_is
3 Replies
4. Solaris
Not sure why solaris couldn't detect the geometry of a hard disk which has a working OS of winxp pro.
Is it due to the different OS that the partition information is stored in different location?
When I type '"format" it is shown as below,
c3d1 < drive type unknown>... (5 Replies)
Discussion started by: just.srad
5 Replies
5. Shell Programming and Scripting
Friends.
I have to compare iostat -x output with a tool on solaris. Now there is column called wait in the output field which is showing zero. Now, in order to create some load on my system this is what i am doing
I am creating a file using dd command , the size of which is... (5 Replies)
Discussion started by: achak01
5 Replies
6. Red Hat
A find for the "iostat" command on a redhat 5 update 4 comes back with no results.
Any separate rpm needs to be installed to get the binary for this ?
Thanks in advance. (1 Reply)
Discussion started by: uxadmin007
1 Replies
7. Shell Programming and Scripting
How to create a symbolic link to a command with certain argument?
When I man hexdump, it is said in the man page that "-C Canonical hex+ASCII display...Calling the command hd implies this option". Actually it is. hd equals to hexdump -C.
And then I examined the ln command but find it is a... (5 Replies)
Discussion started by: vistastar
5 Replies
8. OS X (Apple)
I'm new to playing with the command line on OS X and am puzzled by the response I am getting from the find command. I have a file structure similar to the following
/Volumes/
../Drobo/
../../Pictures/
../../../Image 1/
../../../../Image 1.jpg
../../../../Previews/
../../../../../Image... (2 Replies)
Discussion started by: Denrael
2 Replies
9. AIX
Hello, I support Oracle 11g on AIX 7.1.
Using the command
$iostat -D hdisk2 hdisk4 hdisk5 5
I get the following output:
hdisk5 xfer: %tm_act bps tps bread bwrtn
44.0 1.4M 178.2 1.4M 14.7K
read: ... (3 Replies)
Discussion started by: oracledba1024
3 Replies
10. Programming
Facebook had a mathematics problem which was as thus:-
6/2(1+2) = ?
Answer is 9.
My ancient Casio FX 730P mini computer written exactly as that gives 'error' only.
Now take a look at shell versions, and a python version:-
Last login: Wed Sep 14 18:04:04 on ttys000
AMIGA:barrywalker~>... (6 Replies)
Discussion started by: wisecracker
6 Replies
LEARN ABOUT DEBIAN
mkqdisk
mkqdisk(8) Quorum Disk Management mkqdisk(8) NAME
mkqdisk - Cluster Quorum Disk Utility WARNING
Use of this command can cause the cluster to malfunction. SYNOPSIS
mkqdisk [-?|-h] | [-L] | [-f label] [-c device -l label] [-d [-d ...]] DESCRIPTION
The mkqdisk command is used to create a new quorum disk or display existing quorum disks accessible from a given cluster node. OPTIONS
-c device -l label Initialize a new cluster quorum disk. This will destroy all data on the given device. If a cluster is currently using that device as a quorum disk, the entire cluster will malfunction. Do not run this on an active cluster when qdiskd is running. Only one device on the SAN should ever have the given label; using multiple different devices is currently not supported (it is expected a RAID array is used for quorum disk redundancy). The label can be any textual string up to 127 characters - and is therefore enough space to hold a UUID created with uuidgen(1). -f label Find the cluster quorum disk with the given label and display information about it. -L Display information on all accessible cluster quorum disks. -d Increase debugging level. Specify multiple times for more information. Currently, specifying more than twice has no effect. SEE ALSO
qdisk(5), qdiskd(8), uuidgen(1) July 2006 mkqdisk(8)