After some more research, I think I'm wrong on this. It's still a network
problem, not a SAN issue, but I think you're running
out of ports. Check your time_wait_interval (ndd -get /dev/tcp
tcp_time_wait_interval). If it's 240000, you might want to
decrease it. Also, check netstat output for TIME_WAIT (netstat -a | grep
TIME_WAIT). A lot of those, and you should definitely
decrease the time_wait_interval.
George
On Thu, 20 Nov 2003, George Drew wrote:
> Your configuration is broken. The 126 in parenthesis is system errno 126,
> EADDRNOTAVAIL. It is returned when an attempt is made
> to bind a raw socket to a name. This error is telling you that the name
> you're requesting is not available on this system.
>
> Double-check how you've configured your hardware for netbackup's use. You
> might want to investigate the REQUIRED_INTERFACE
> vm.conf entry. For troubelshooting, remove any SERVER entries from vm.conf,
> if any exist.
>
> George
>
> On Wed, 19 Nov 2003, Robert Loh wrote:
>
> > Hi Marianne,
> >
> > Thanks for your infor, I only have a single sun server as backup and master
> > server going to Compaq SAN switch with two JNI cards, two fibre conneciton
> > to the switch, the first JNI card can see the robotic arm and the first 8
> > tape drive, the second JNI card see the other 8 tape drives. JNI card was
> > configured persistent binding by ports of the Tape library fibre-router, the
> > ATL FC230 fibre router has three ports, the frist two ports were used as the
> > last port is for failover. The first port going to two scsi bus, the first
> > scsi bus to robotic arm and two drives, second scsi bus to another two
> > drives, so each scsi port mapped to two drives. The setup is Fabric SAN
> >
> > I still have a gut feel that below error messages is somethings to do with
> > NetBackup port communication or may be network related but don't know where
> > should I look into. tldd is the middleman talking to ltid and tldcd, somehow
> > tldd couldn't talk to tldcd and timeout then tape library going to DOWN
> > state, suprisingly after that tldd manages to talk to tldcd again and tape
> > libary going UP state again, it keeps on and on, and how this thing happen
> > and how to prevent it ? why tldd couldn't talk to tldcd in what kind of
> > circumstances ?
> > Nov 16 17:29:18 sgs15i-001 tldd[21598]: TLD(2) [21598] unable to connect to
> > tldcd on sgs15i-001: Cannot assign requested address (126)
> >
> > Nov 16 17:29:18 sgs15i-001 tldd[1322]: TLD(2) going to DOWN state, status:
> > Control daemon connect or protocol error
> > Nov 16 17:33:22 sgs15i-001 tldd[1322]: TLD(2) going to UP state
> >
> > Thanks a lot
> > Robert
> >
> > -----Original Message-----
> > From: Marianne van den Berg
> > [mailto:Marianne.vandenBerg AT lechabilestorage.co DOT za]
> > Sent: Tuesday, November 18, 2003 8:00 PM
> > To: Robert Loh
> > Subject: RE: [Veritas-bu] Cannot assign requested address ?
> >
> >
> > Hi Robert
> >
> > Are you by any chance using Arbitrated Loop in your SAN?
> >
> > Have a look at the extract from the SSO manual - it looks as if Arbitrated
> > Loop gives you the worst performance.
> >
> > Fibre Channel Topologies
> >
> > Fibre channel protocol is designed to be low in overhead so it can function
> > over a device channel, yet flexible so messages have addresses and can be
> > efficiently switched. Devices that participate in fibre channel networking
> > have names and identities. Fibre channel was designed to be as adept for
> > carrying storage protocols (SCSI, HIPPI, and IPI) as carrying network
> > protocols (IP and ATM).
> >
> > At the lowest layers of the 5-layer protocol, fibre channel defines the
> > following types of device connections.
> >
> > Point-to-Point
> >
> > This topology is the closest to a SCSI connection. It consists of two fibre
> > channel devices connected directly together. The transmit fibre of one
> > device goes to the receive fibre of the other device, and vice versa. There
> > is no sharing of the media, so the devices enjoy the total bandwidth of the
> > link. A link initialization is required of the two devices before
> > communication can begin.
> >
> > Arbitrated Loop (FC-AL)
> >
> > This is the most dominant fibre channel topology because of low cost, but is
> > falling out of favor because of poor performance. This topology allows up
> > to 126 devices to reside on a physical loop.Any device on the loop must
> > negotiate access to the loop in order to transfer data. The media is shared
> > among the devices, limiting each device's access.
> >
> > Arbitrated loops are private in that devices on the loop are only visible to
> > the host where the loop is connected and only a minimal self-identification
> > procedure is conducted at system startup.
> >
> > Switched Fibre Channel (Fibre Channel Fabric or Fabric)
> >
> > This topology is used to connect many (up to 224) devices in a cross-point
> > switched configuration. The benefit of this topology is that many devices
> > can communicate at the same time and the media can be shared. It also
> > requires fibre channel switches.
> >
> >
> >
> >
> >
> >
> >
> >
> > Marianne van den Berg
> > Senior Support Engineer
> > LECHABILE Storage Solutions (PTY) Ltd
> > Tel: +27 11 519 8134
> > Mobile: +27 82 468 8107
> > mailto: Mariannev@LechabileSS <mailto:Mariannev AT LechabileSS.co DOT za>
> > .co.za
> >
> > This e-mail and attachments relating thereto, is intended for the
> > above-mentioned recipient. If you have received this e-mail in error, kindly
> > notify the sender and delete it immediately as it contains information
> > relating to the official business of Lechabile Storage Solutions (Pty)
> > Limited, which is confidential, legally privileged and proprietary to
> > Lechabile Storage Solutions (Pty) Limited. Lechabile Storage Solutions (Pty)
> > Limited does not own and endorse any other content. The views and opinions
> > expressed in this e-mail are those of the sender unless clearly stated as
> > being those of Lechabile Storage Solutions (Pty) Limited.
> >
> > -----Original Message-----
> > From: Robert Loh [mailto:robert AT softsource.com DOT sg]
> > Sent: Tuesday, November 18, 2003 3:52 AM
> > To: Marianne van den Berg
> > Subject: RE: [Veritas-bu] Cannot assign requested address ?
> >
> >
> > Hi Marianne,
> >
> > I am very sure that the device configuration files are all correct, if not I
> > would be able to mount/unmount backup to respective drives successfully. We
> > are using JNI so persistent binding was setup and below error messages
> > occurred no because of server after reboot.
> >
> > Weirld thing is it says going to down then after few minutes it come up by
> > itself. DOWN and UP make me feel very worry. Any other thought ? :)
> > Oh yes, sgscan return the right config.
> > CHeers....
> > Robert
> >
> > -----Original Message-----
> > From: Marianne van den Berg
> > [mailto:Marianne.vandenBerg AT lechabilestorage.co DOT za]
> > Sent: Monday, November 17, 2003 4:50 PM
> > To: Robert Loh
> > Subject: RE: [Veritas-bu] Cannot assign requested address ?
> >
> >
> >
> > Hi Robert
> >
> >
> > Have you checked output from sgscan? Compare this output with your device
> > configuration.
> >
> > If persistent binding is not in place, it is possible that device names were
> > changed after a reboot.
> >
> > Regards
> >
> > Marianne van den Berg
> > Senior Support Engineer
> > LECHABILE Storage Solutions (PTY) Ltd
> > Tel: +27 11 519 8134
> > Mobile: +27 82 468 8107
> > mailto: Mariannev@LechabileSS <mailto:Mariannev AT LechabileSS.co DOT za>
> > .co.za
> >
> > This e-mail and attachments relating thereto, is intended for the
> > above-mentioned recipient. If you have received this e-mail in error, kindly
> > notify the sender and delete it immediately as it contains information
> > relating to the official business of Lechabile Storage Solutions (Pty)
> > Limited, which is confidential, legally privileged and proprietary to
> > Lechabile Storage Solutions (Pty) Limited. Lechabile Storage Solutions (Pty)
> > Limited does not own and endorse any other content. The views and opinions
> > expressed in this e-mail are those of the sender unless clearly stated as
> > being those of Lechabile Storage Solutions (Pty) Limited.
> >
> > -----Original Message-----
> > From: Robert Loh [mailto:robert AT softsource.com DOT sg]
> > Sent: Monday, November 17, 2003 10:04 AM
> > To: veritas-bu AT mailman.eng.auburn DOT edu
> > Subject: [Veritas-bu] Cannot assign requested address ?
> >
> >
> >
> > Hi folks,
> >
> > NetBackup 3.4.2 with ATL P7000 16 x SDLT220 LVD drives connected to Compaq
> > SAN switch with a single backup server (media server)
> >
> > Always receiving error below in my /var/adm/messages in Solaris 8
> >
> > Nov 16 17:29:18 sgs15i-001 tldd[21598]: TLD(2) [21598] unable to connect to
> > tldcd on sgs15i-001: Cannot assign requested address (126)
> >
> > Nov 16 17:29:18 sgs15i-001 tldd[1322]: TLD(2) going to DOWN state, status:
> > Control daemon connect or protocol error
> > Nov 16 17:33:22 sgs15i-001 tldd[1322]: TLD(2) going to UP state
> > Nov 16 18:15:53 sgs15i-001 tldd[9349]: TLD(2) [9349] unable to connect to
> > tldcd on sgs15i-001: Cannot assign requested address (126)
> >
> > Nov 16 18:15:53 sgs15i-001 tldd[1322]: TLD(2) going to DOWN state, status:
> > Control daemon connect or protocol error
> > Nov 16 18:18:22 sgs15i-001 tldd[1322]: TLD(2) going to UP state
> >
> > NetBackup error or tape library error !
> > Appreciate for helping.
> > Thanks.Robert
> >
> >
> >
> >
> >
> >
>
|
