[3310] ssmon can generate ASC: 0x29 (power on, reset, or bus reset occurred)

Asset ID:	1-72-1007691.1
Update Date:	2018-01-03
Keywords:

Solution Type Problem Resolution Sure

Solution 1007691.1 : [3310] ssmon can generate ASC: 0x29 (power on, reset, or bus reset occurred)

Applies to:

Sun Storage 3310 Array - Version Not Applicable and later
Sun Storage 3320 SCSI Array - Version Not Applicable and later
All Platforms

Symptoms

If you have the following messages in the host messages, 3 times a day and more when the I/O activity increases, they can be generated by ssmon daemon or ssconsole GUI (graphical user interface) when "in-band" monitoring is configured.

Jan 3 08:02:59 bretagne scsi: [ID 107833 kern.warning] WARNING: /pci@1d,700000/pci@2/scsi@5/sd@1,4 (sd154):
Jan 3 08:02:59 bretagne Error for Command: read Error Level: Retryable
Jan 3 08:02:59 bretagne scsi: [ID 107833 kern.notice] Requested Block: 533456 Error Block: 533456
Jan 3 08:02:59 bretagne scsi: [ID 107833 kern.notice] Vendor: SUN Serial Number: 25AFBEBB-01

Jan 3 08:02:59 bretagne scsi: [ID 107833 kern.notice] Sense Key: Unit Attention
Jan 3 08:02:59 bretagne scsi: [ID 107833 kern.notice] ASC: 0x29 (power on, reset, or bus reset occurred), ASCQ: 0x0, FRU: 0
x0
Jan 3 11:16:33 bretagne scsi: [ID 107833 kern.warning] WARNING: /pci@1d,700000/pci@2/scsi@5/sd@1,4 (sd154):
Jan 3 11:16:33 bretagne Error for Command: read(10) Error Level: Retryable
Jan 3 11:16:33 bretagne scsi: [ID 107833 kern.notice] Requested Block: 136343840 Error Block: 136343840
Jan 3 11:16:33 bretagne scsi: [ID 107833 kern.notice] Vendor: SUN Serial Number: 25AFBEBB-01

Jan 3 11:16:33 bretagne scsi: [ID 107833 kern.notice] Sense Key: Unit Attention
Jan 3 11:16:33 bretagne scsi: [ID 107833 kern.notice] ASC: 0x29 (power on, reset, or bus reset occurred), ASCQ: 0x0, FRU: 0
x0
Jan 3 14:28:23 bretagne write: [ID 201458 auth.crit] /dev/dtremote in utmpx is not a tty
Jan 3 14:48:56 bretagne su: [ID 810491 auth.crit] 'su root' failed for 98000 on /dev/pts/9
Jan 3 22:36:59 bretagne scsi: [ID 107833 kern.warning] WARNING: /pci@1d,700000/pci@2/scsi@5

Jan 3 22:36:59 bretagne Error for Command: read(10) Error Level: Retryable
Jan 3 22:36:59 bretagne scsi: [ID 107833 kern.notice] Requested Block: 139878414 Error Block: 139878414
Jan 3 22:36:59 bretagne scsi: [ID 107833 kern.notice] Vendor: SUN Serial Number: 25AFBEBB-01

Jan 3 22:36:59 bretagne scsi: [ID 107833 kern.notice] Sense Key: Unit Attention
Jan 3 22:36:59 bretagne scsi: [ID 107833 kern.notice] ASC: 0x29 (power on, reset, or bus reset occurred), ASCQ: 0x0, FRU: 0

Notice : "in-band" means ssmon or ssconsole communicate through the SCSI connection, vs out-of-band through the network connection

Cause

ssmon active using "in band" access to the storage

Solution

Before planning any hardware replacement, it's compulsory to apply the minimum
level of patch for the I/O drivers on the host, and upgrade the firmware of the array controllers and
internal disks before verifing the ssmon configuration as described below:

1) Verify software/firmware:

o All array firmwares should be at the latest revisions: 4.23C for 3310, and 4.21E for 3320 Controller firmware (minimum requirement),
SAF-TE firmware rev 1180 for SE3310, and 1185 for the SE3320 (minimum requirement).
o All HBA, and SCSI drivers should be up-to-date for the particular Solaris release.
Use an available diagnostic tool such as explominer, or patchdiag, to obtain the required patches to install.
o Ensure SSCS (ssconsole) is at the latest version: 2.5 (minimum requirement).

2) Check if the ssagent is configured "out-of-band":

Edit the ssagent configuration file: /var/opt/SUNWsscs/ssagent/ssagent.cfg

o Verify the value of variable PRIAGENT_OVER_INBAND

A value of 0 for PRIAGENT_OVER_INBAND indicates that the direct "in-band" scsi
access method is selected. This is the default value. If this value is set to 0, proceed to the next step if you wish to configure ssagent "out-of-band".

3) How to configure ssagent "out-of-band":

When the network interface of the SE3310 or SE3320 is configured to a LAN.

There are two methods:

o By command line:

a) Stop the ssagent:

/etc/init.d/ssagent stop

b)Edit the file: /var/opt/SUNWsscs/ssagent/ssagent.cfg
and set to 1 the value of variable PRIAGENT_OVER_INBAND.

PRIAGENT_OVER_INBAND=1

c) Check that the 2 following lines exist for each array in this file:

PRIAGENT_IP=xxx.xxx.xxx.xxx

PRIAGENT_PORT=58632

d) Verify the network interface of [3310] is correctly configured in the ssagent.cfg file:

PRIAGENT_IP=xxx.xxx.xxx.xxx ( IP address of [3310] )

PRIAGENT_PORT=58632 ( address port by default )

e)Start the ssagent:

/etc/init.d/ssagent start

o By the GUI (sscs):

Follow the steps described in the Sun Storedge 3000 Family Configuration
Services 2.5 User's Guide (817-3337-18) in the section: "To Use In-Band Management".

Relief/Workaround

If "in-band" monitoring is the only alternative for a particular array, you can stop the ssagent
software:

/etc/init.d/ssagent stop

and monitor for one week or more depending upon the frequency of the messages,
to see if they stop.

If the messages still occur, it's advised to place the array on a private network
and apply the above step 3.

Otherwise, if errors still occur, the cause is more complex and will need additional troubleshooting.

For out-of-band it's recommended to configure the network interface of the array
on a private network and apply the above step 3.
This recommendation is to prevent network port scanning software interfering with
the TCP/IP stack on the array.

Attachments

This solution has no attachment