Filer General

Filer General

Messages at screen is configured at

/etc/syslog.conf.sample

By default, there is no such file, but if user modifies this file, they will have

/etc/syslog.conf ----------- which will tell where to direct

messages at screen ( typically

/etc/messages )

Sysconfig –t ( tape information )

Source –v /etc/rc - this command reads and executes any file

containing filer commands line by line

Auto support ( user – trigger support )

Options.autosupport.doit autosupport@netapp.com

Telneting to Filer

Only one user can do telnet

Options telnet

Autosupport Configuration

Filer>Options autosupport

autosupport.mailhost < >

autosupport.support.to < autosupport@netapp.com >

autosupport.doit <string>

autosupport.support.transport https or smtp

autosupport.support.url < url address must be reachable >

Autosupport troubleshooting

1. ping netapp.com from filer

2. TCP 443 SSL should be open at SMTP server

SMTP server may stay in DMZ side

3. Mail relay in exchange must be specified. Filer’s host name or IP address must be specified in mail relay. Routing for netapp.com or routing by this host or routing by this ip must be enabled for filer. Filer is acting as a SMTP client. In general setup of mail system, no SMTP client is able to send the mail thru mail server to other SMTP server when host’s identity is different as far as mail id is concerned. Relaying is blocked generally.

4. Proxy server http / https must pass http url

Raid Scrub weekly

a. raid.scrub.duration 360

b raid.scrub.schedule sun@01

a. scrub to happen for only 6 hrs

b. forcing the scrub to start on Sunday at 1 am

RAID group

Vol add vol0 –g rg0 2 add 2 disks to raid group 0 of vol0

Vol options yervol raidsize 16 changes the raidsize settings

of the vol yervol to 16

vol create newvol –t raid_dp –r 16 32@36

- newvol creation with raid_dp protection.

RAID group size is 16disks. Since the vol

consists of 32 disks, those will form 2 RAID

group, rg0 & rg1

Max Raid groupsize

Raid DP 28

Raid 4 14

vol options for snapshots

nosnapdir off < default off >

nosnap off < default off >

Disk Fail/unfail

priv set advanced when disk goes bad

disk fail partially then prefail copy

disk unfail is seen when sysconfig -r

sysconfig –d is done. Somestimes it may

Disk troubleshoot just hang there, so disk fail

Priv set advanced -i <disk name> would

Led_on < 1d.16> release the disk &

Led_off < drive id > reconstruct the the RAID

group

Blink_on 4.19 ( failed disk now will be orange )

Blink_off 4.19

Spare disks in vol

Vol status -s

FAS 270 ( this must be done, otherwise not seen )

Priv set advanced

Disk show –v ----- to see who owned it. If this has come from

Another filer, disk block header needs to

Remove. For that

Disk unfail <disk id>

Disk assign 0b.23

Fcadmin device_map

If drive not shown in filer view

Filer> storage show disk -p

Zeroing disks

Priv set advanced

Disk zero spares --- to zero out the data in all spares

Sysconfig –r ( will show % of zero disk ) - spares disks

R100 & R150 Disk Swap

1. find bad disk , identify it

2. type disk swap < disk id >

3. Remove disk

4. Wait 20 sec

5. disk swap again

6.insert new disk

7. wait 20 sec to rescan

Out of inodes

Filer General NFS CIFS SnapMirror SnapVault NDMPCopy SM SQL Search

1. Check % used of inodes by

Filer> df –i

2. to increase

Filer> maxfiles < vol name > <max>

NVRAM

Battery check

Filer> priv set diag

Filer> nv

=> should show battery status as OK and voltage as

NVRAM3 6V

Raidtime out in options raid controlls ( 24 hr ) the trigger when bat low

In 940s – NVRAM5 is used as Cluster interconnect card as well, “two in one” on slot 11

Time Deamond

(port 123, 13, 37 must be open)

When there is large skew, lot of messages from

CfTimeDaemon : displacements /skews:10/3670,10/3670, 11/3670

Because of this hourly snapshot creation also fails or in progress message appears.

Because of timed.max_skew set to 30 min, we may see above message in every 30min- 1 hr

If we set this to 5s and see how skew happening – if we see lot of skew messages (once we turned ON to timed.log ON ), MB replacement may require.

For temporary do

Cf.timed.enable ON on both cluster filers and watch those off errors

Checking from unix host

# ntptrace –v filername

From filer check

Filer>options timed

( check all the options of this )

From filer view => set date and time : Synchronize now < ip of NTP server > => do synchronize now and check NTP from unix host.

Tip : if there are multiple interfaces in filer, make sure that they are properly listed in NIS or DNS server – same host name , multiple ip addresses may require

BPS ( Blocks Per Sector ) of Disk

Block Append Checksum requires each disk drive to format it to 520 or 512 BPS per sector This provides a total of 4160 bytes in 8 sectors. This space is broken into two parts. First part is 4096 bytes ( 4K - the WAFL file size ) of file system data. The remaining 64 bytes contain the checksum data for previous 4096 bytes. In this manner, the checksum block is appeneded to each block of data.

Enviromental Status

The top line in each channel says failures to yes , if there is any.

Subsequent messages should say

Power

Cooling

Temperature

Embedded switching [ all to none ]

( if there is no problem )

Volume

Vol options vol0

Vol status vol0 -r ( raid info of volume )

Sysconfig –r

Vol options vol0 raidsize 9

Vol add vol0 <number of disk >

Vol status –l ---- to display all volume

Aggr Volume creation

Filer> create aggr1 10

Filer> vol create log1 aggr1 20g

When vol is gone bad

Vol wafliron start <vol name> -f

To list broken disk in volume

Vol status –f

Sysconfig –r will tell the failed disks

Double Parity

Vol create –t raiddp –r 2 ( minimum of two )

(there are two parity disks for holding parity and double parity data)

Enviroment status – like temp/shelf issues

Environment chesis list_sensors

Environment dump

RSH options - rsh access of fier

Options rsh.enabled on

Adminhost needed to add to do RSH ( can be done from filer

View ) - not root. RSH sec settings must be set with either ip or hostname, but with matching username for logon accounts ( not root, but the domain admin account )

RSH access from unix host

# rsh –l root <console p/w> <ip of filer> “<command>”

( add this unix host in /etc/hosts.equiv file – similar for windows host as well )

( this command can be corned in unix to make it scheduled )

RSH Port 514 / TCP

Registry Walk

Filer> registry walk status.vol.<vol name>