The V240 is one system that you DO NOT make static on either the server side
(setting the NIC) nor on the switch side.
Any Sun system with the BGE interface has to run at Auto/Auto on both sides.
Sun has documented this in several Tech Notes.
I am running 15 V240's as Masters and/or Media servers. I have checked and
rechecked this and can you it works on Auto/Auto.
By the way, on the V240, the Network Admin console has to have its Switch
port set to 10/Half to work correctly.
Bobby Williams
2205 Peterson Drive
Chattanooga, Tennessee 37421
423-296-8200
-----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 Priddy,
John
Sent: Friday, October 14, 2005 1:32 AM
To: Veritas List
Subject: RE: [Veritas-bu] how to deal with drive shoe-shining
You say from 'local disk to tape' im assuming you don't mean that the
volumes on the netbackup server itself are backing up slow, but the network.
If your talking about remote hosts over the network, (and assuming were
talking solaris) then you should check to be sure 100mb full is static on
both the host and switch as someone mentioned before. To give some more
detail as im resolving one of these issues right now....
On the host/netbackup server /etc/system should contain a line like (this
particular one is a v240):
--
* Setting up Autonegotiation off and Full Duplex 100MBPS ON set
bge0:bge_adv_autoneg_cap=0 set bge0:bge_adv_100T4_cap=0 set
bge0:bge_adv_100fdx_cap=1 set bge0:bge_adv_100hdx_cap=0 set
bge0:bge_adv_10fdx_cap=0 set bge0:bge_adv_10hdx_cap=0
--
On cisco side you should see lines in the running-config for the specific
interface like (this is the 9th plate/port 48 on a 6509):
--
!
interface FastEthernet9/48
no ip address
duplex full
speed 100
switchport
switchport access vlan XX
--
If you don't have your unix/cisco nic environment static I highly recommend
you do so, as 9 out of 10 problems I have had with speed/bandwith between
cisco and sun have been resolved setting these to static on both ends.
Interestingly enough, only when we decided to get rid of legato and move to
netbackup did I even notice that we had these network issues. I guess you
could say netbackup is a good network performance tool. Why cisco and sun
cant play nice together with regards to auto negotiation is beyond me,
though were running a pretty old IOS at my shop.
All that being said, I also use an L20, but w/ 2 DLT drives on solaris 10,
and I get peak performance with the following: NUMBER_DATA_BUFFERS 16,
SIZE_DATA_BUFFERS 65536. All of this was set as per VTS performance
documentation I found online. I rarely fall below the max capabilities of
the drives with these settings, though when I do it's almost certainly my
horribly old hardware I'm grabbing it from.
If all that doesn't help, what are the specs of your media server?
Maybe it can't hang with the traffic coming in? Have you done tests
transferring large amounts of data between the machines you're having
problems with? If its only one particular machine, it may sound silly, but
check the network cable path, I recently had sporadic nfs issues because of
a network wire that was wrapped around 2 emc power wires.
You never know what's below the tiles in a data center.
-----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 Charles
Ballowe
Sent: Thursday, October 13, 2005 5:59 PM
To: Anthony Tocco
Cc: Zhisong Jin; Veritas List
Subject: Re: [Veritas-bu] how to deal with drive shoe-shining
There is no MP5 for 5.1 -- there's MP3A which you should be running, + the
patches from yesterday.
-Charlie
On 10/13/05, Anthony Tocco <ttocco AT colltech DOT com> wrote:
> Have you checked to see if you are forcing your nic's and switches.
If
> they are set to auto negotiate then that could b e your issue. Also
> upgrade it to MP5 there is network issues with mp3
>
> Tony Tocco
> ttocco AT colltech DOT com
>
>
> -----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 Zhisong
> Jin
> Sent: Thursday, October 13, 2005 4:28 PM
> To: Veritas List
> Subject: [Veritas-bu] how to deal with drive shoe-shining
>
> I have a SUN L20 tape library runing as a media server for veritas
> netbackup 5.1 on SUN solaris 9. the systems runs however SOMETIMES
> performance is terrible.
>
> for example, a recent backup from a local disk to tape got 60KB/s in
> throughput.
>
> enviroment : solaris 9, netbackup 5.1 MP3, drive cleaned regularly.
>
> on media server /etc/system:
> set shmsys:shminfo_shmmax=4294967295
> set shmsys:shminfo_shmmin=1
> set shmsys:shminfo_shmseg=10
> set shmsys:shminfo_shmmni=512
> set semsys:seminfo_semmns=2048
> set semsys:seminfo_semmni=2048
> set semsys:seminfo_semmsl=250
> set msgsys:msginfo_msgmni=1024
> set msgsys:msginfo_msgtql=1024
> set semsys:seminfo_semopm=128
>
> on media server /usr/openv/netbackup/
> NET_BUFFER_SZ = 262144 in /usr/openv/netbackup
> SIZE_DATA_BUFFERS = 262144 in /usr/openv/netbackup/db/config
> NUMBER_DATA_BUFFERS = 32;
>
> I even have an entry per veritis recommedation
> SIZE_DATA_BUFFERS_DISK = 1048576
> NUMBER_DATA_BUFFERS_DISK= 16
>
> can anybody offer some clues what is wrong here?
>
> Thanks.
> Jason
>
> _______________________________________________
> Veritas-bu maillist - Veritas-bu AT mailman.eng.auburn DOT edu
> http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
>
> _______________________________________________
> Veritas-bu maillist - Veritas-bu AT mailman.eng.auburn DOT edu
> http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
>
_______________________________________________
Veritas-bu maillist - Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
_______________________________________________
Veritas-bu maillist - Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
|
