Veritas-bu

[Veritas-bu] Backups slow to a crawl

2005-03-23 10:56:22
Subject: [Veritas-bu] Backups slow to a crawl
From: Bill_Jorgensen AT csgsystems DOT com (Jorgensen, Bill)
Date: Wed, 23 Mar 2005 08:56:22 -0700
Jeff:

A few things to consider (assuming a Sun sever as the NBU master):

1.) Are you aware of anything that has changed on your NBU server?
2.) Are you aware of anything that has changed with your network?
(Providing you are doing Ethernet-based backups. If not, what about the
SAN?)
3.) Are you aware of any changes to the policies?

If no to the above try the following:

1.) Find out what Veritas recommends for your environment for these two
variables:
NUMBER_DATA_BUFFERS
SIZE_DATA_BUFFERS
These are found in /usr/openv/netbackup/db/config. They may not give
them to you if you open a ticket with the solution center (Professional
Services). Ask around if they do not.

2.) Check the network driver settings for a few things. This depends on
the network type you are using. 100Mb-switched, 10Mb-switched, etc.

root[prod-backup:/]# ndd -get /dev/qfe adv_autoneg_cap
1
root[prod-backup:/]# ndd -get /dev/qfe adv_100hdx_cap
1
root[prod-backup:/]# ndd -get /dev/qfe adv_100fdx_cap
1
What the output above is stating is that the qfe driver is set at 100
half and full duplex, and autonegotiate. Once you know how the network
driver is configured go to your network guys and ask them to see how the
port on the switch is configured (unless you are the network guy). If
the port is NOT set to 100-full or autonegotiate have them set it
accordingly.

3.) Reseat the RJ-45 connectors for the physical connections.

These are some things that have bit us in the past.

Good luck,

Bill

--------------------------------------------------------
     Bill Jorgensen
     CSG Systems, Inc.
     (w) 303.200.3282
     (p) 303.947.9733
--------------------------------------------------------
     UNIX... Spoken with hushed and
     reverent tones.
--------------------------------------------------------

-----Original Message-----
From: veritas-bu-admin AT mailman.eng.auburn DOT edu
[mailto:veritas-bu-admin AT mailman.eng.auburn DOT edu] On Behalf Of Jeff
McCombs
Sent: Wednesday, March 23, 2005 6:51 AM
To: veritas-bu AT mailman.eng.auburn DOT edu
Subject: [Veritas-bu] Backups slow to a crawl

Gurus,

    NB 5.0 MP4, single combination media/master server, Solaris 9.
Overland
Neo 2000 26-slot 2 drive DLT.

    I'm noticing that for some reason or another, all of my client
backups
have slowed to a _crawl_. A _cumulative_ (!) backup of local disk on a
Sun
V100 is taking somewhere on the order of 2 hours at this point, and with
over 40 systems, I'm blowing past my window  consistently.

    I'm not quite sure what's going on here, but as I sit and watch the
output from 'iostat', I'm noticing that rmt/1 (the 2nd drive in the Neo)
is
fluxuating between 100% busy, with kw/s at close to zero, and busy @
1-15%
and kw/s up into the 1000's.

    rmt/0 seems to be fine, kw/s sits consistently up in the 1.8-2K
range,
while busy is anywhere from 2% - 30% on average. My other disks aren't
working hard, CPU isn't loaded and I've got plenty of memory.

    The policy I'm using allows for multiple datastreams, no limits on
jobs,
and most schedules allow for an MPX of 2. I'm backing up
ALL_LOCAL_DRIVES on
all clients, and I'm not using any NEW_STREAM directives. I'm not seeing
any
errors on the media either.

    Can anyone shed some light on what might be happening here? Am I
looking
at a drive that might be having some problems, or am I barking up the
wrong
tree, and it's something else entirely?

    A small sample of iostat output covering the affected devices is
below.

sample (extra disks removed from putput);
root@backup(pts/1):~# iostat -nx 1 100
                    extended device statistics
    r/s    w/s   kr/s   kw/s wait actv wsvc_t asvc_t  %w  %b device
    0.0    4.1    0.0  252.2  0.0  0.0    0.0    5.9   0   2 rmt/0
    0.0    4.6    0.0  278.4  0.0  0.1    0.0   27.3   0  12 rmt/1

    0.0    4.1    0.0  252.3  0.0  0.0    0.0    5.9   0   2 rmt/0
    0.0    4.6    0.0  278.4  0.0  0.1    0.0   27.3   0  12 rmt/1

    0.0   33.0    0.0 2076.4  0.0  0.2    0.0    5.8   0  19 rmt/0
    0.0    2.0    0.0  125.8  0.0  1.0    0.0  490.0   0  98 rmt/1

    0.0   38.0    0.0 2394.0  0.0  0.2    0.0    5.4   0  21 rmt/0
    0.0    8.0    0.0  504.0  0.0  1.0    0.0  124.9   0 100 rmt/1

    0.0   27.0    0.0 1701.1  0.0  0.2    0.0    6.5   0  17 rmt/0
    0.0    2.0    0.0  126.0  0.0  1.0    0.0  499.9   0 100 rmt/1

    0.0   33.0    0.0 2078.9  0.0  0.2    0.0    5.3   0  18 rmt/0
    0.0    0.0    0.0    0.0  0.0  1.0    0.0    0.0   0 100 rmt/1

    0.0   16.0    0.0 1008.0  0.0  0.1    0.0    6.2   0  10 rmt/0
    0.0   13.0    0.0  819.0  0.0  0.6    0.0   48.4   0  63 rmt/1

    0.0   40.0    0.0 2520.1  0.0  0.2    0.0    5.9   0  24 rmt/0
    0.0    0.0    0.0    0.0  0.0  1.0    0.0    0.0   0 100 rmt/1

    0.0   33.0    0.0 2078.9  0.0  0.2    0.0    5.3   0  18 rmt/0
    0.0   10.0    0.0  630.0  0.0  1.0    0.0   99.9   0 100 rmt/1


-- 
Jeff McCombs                 |                                    NIC,
Inc
Systems Administrator        |
http://www.nicusa.com
jeffm AT nicusa DOT com             |                                NASDAQ:
EGOV
Phone: (703) 909-3277        |        "NIC - the People Behind
eGovernment"
--
    "From space, I saw the fragile blue earth without borders."
        - Onizuka.


_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu