HPE Community read-only access December 15, 2018
This is a maintenance upgrade. You will be able to read articles and posts, but not post or reply.
Hours:
Dec 15, 4:00 am to 10:00 am UTC
Dec 14, 10:00 pm CST to Dec 15, 4:00 am CST
Dec 14, 8:00 pm PST to Dec 15, 2:00 am PST
Operating System - OpenVMS
cancel
Showing results for 
Search instead for 
Did you mean: 

Does HSJ Disks share CI paths (Two Paths)

 
Shailender_1
Advisor

Does HSJ Disks share CI paths (Two Paths)

I have a small query here. Just wanted to know whether HSJ50 Storage disk connetced through two CI paths do load balancing(Share Traffic) between the path or only best path is selected and continue to serve through that path only.

How do I check which path the disk is being served and the traffic

Thanks
10 REPLIES
Hein van den Heuvel
Honored Contributor

Re: Does HSJ Disks share CI paths (Two Paths)

It only uses one at a time per disk.

You can use "SET PREFERRED" to make sure the load is balanced.

http://h71000.www7.hp.com/doc/731FINAL/4477/4477pro_009.html

Details or restrictions may exist depending on the exact OpenVMS version being used, the systems and adapters.

If the CI performance is at all a concern to you, then please consider alternative, more recent (from this century), storage solutions. Per disk space has increased 100x since the CI devices, Disk access times have improved 2x - 4x, Disk transfer times have increase 10x.

Hope this helps some,
Hein van den Heuvel (at gmail dot com)
HvdH Performance Consulting


Robert V Dixon
Occasional Advisor

Re: Does HSJ Disks share CI paths (Two Paths)

>> How do I check which path the disk is being served...

$SH DEV D/FULL should tell you the primary path/controller

Regards
Rob
Shailender_1
Advisor

Re: Does HSJ Disks share CI paths (Two Paths)

Thanks for all your help.

Show dev/full tells info about primary and secondary controllers not the CI path. I am intrested in knowing the CI path being used for disk serving for a perticular disk and the traffic count. Is it possible to get this info?

Regards

Disk $5$DUA800: (HSJ03), device type MSCP served SCSI disk, is online, mounted,
file-oriented device, shareable, served to cluster via MSCP Server, error
logging is enabled.

Error count 0 Operations completed 1667619
Owner process "" Owner UIC [SYSTEM]
Owner process ID 00000000 Dev Prot S:RWPL,O:RWPL,G:R,W
Reference count 1 Default buffer size 512
Current preferred CPU Id 1 Fastpath 1
Total blocks 17768677 Sectors per track 64
Total cylinders 34705 Tracks per cylinder 8
Host name "HSJ03" Host type, avail HSJ5, yes
Alternate host name "HSJ04" Alt. type, avail HSJ5, yes
Allocation class 5

Volume label "SYSTOOLS" Relative volume number 0
Cluster size 18 Transaction count 1
Free blocks 5300406 Maximum files allowed 467596
Extend quantity 5 Mount count 2
Mount status System Cache name "_DSA0:XQPCACHE"
Extent cache size 64 Maximum blocks in extent cache 530040
File ID cache size 64 Blocks in extent cache 386496
Quota cache size 0 Maximum buffers in FCP cache 4318
Volume owner UIC [SYSTEM] Vol Prot S:RWCD,O:RWCD,G:RWCD,W:RWCD

Volume Status: ODS-2, subject to mount verification, file high-water marking,
write-back caching enabled.
Volume is also mounted on SMP01.
Martin Hughes
Regular Advisor

Re: Does HSJ Disks share CI paths (Two Paths)

$ WRITE SYS$OUTPUT F$GETDVI("$1$DUA0","HOST_NAME")

or

$ WRITE SYS$OUTPUT F$GETDVI("$1$DUA0:","ALT_HOST_NAME")

If you want to check traffic on the controller, run VTDPY.
For the fashion of Minas Tirith was such that it was built on seven levels, each delved into a hill, and about each was set a wall, and in each wall was a gate. (J.R.R. Tolkien). Quote stolen from VAX/VMS IDSM 5.2
Robert V Dixon
Occasional Advisor

Re: Does HSJ Disks share CI paths (Two Paths)

You are quite right. Show dev/full tells info about primary and secondary controllers not the CI path. I misunderstood your question ( as did others, I believe).

Without questioning the validity of your request could I ask how you are planning to use the information about which wire(s) out of the four is carrying traffic and for more details about your configuration?

Regards
Rob
Shailender_1
Advisor

Re: Does HSJ Disks share CI paths (Two Paths)

CI adapters are recognised as PNA and PNB devices

$SMP02/IT,E14249> sho dev pn/full

Device PNA0:, device type CIPCA, is online, shareable, error logging is enabled.

Error count 4 Operations completed 0
Owner process "" Owner UIC [SYSTEM]
Owner process ID 00000000 Dev Prot S:RWPL,O:RWPL,G,W
Reference count 0 Default buffer size 0
Current preferred CPU Id 1 Fastpath 1
Current Interrupt CPU Id 1

Device PNB0:, device type CIPCA, is online, shareable, error logging is enabled.

Error count 3 Operations completed 0
Owner process "" Owner UIC [SYSTEM]
Owner process ID 00000000 Dev Prot S:RWPL,O:RWPL,G,W
Reference count 0 Default buffer size 0
Current preferred CPU Id 0 Fastpath 1
Current Interrupt CPU Id 0

CI throuput is 70Mbps(Mega bits /sec) single path. So if devices are served on single path I should not expect data throughput of more than 6-7 MBPS. If disks are load balanced between the two CI paths I can expect data throuput 140(70 x 2 CI paths)

I generated ECP(Enterprise capacity planner) report. report shows that all the disks are being served through one CI path only. That's why wanted to know How CI works.

Robert V Dixon
Occasional Advisor

Re: Does HSJ Disks share CI paths (Two Paths)

> CI adapters are recognised as PNA and PNB devices
Agreed: Two physical cards seen by VMS as PNA and PNB.

As you say, CI throughput per path is 70Mbps(Mega bits /sec). CIPCA is capable of running dual path so you should see a theoretical max of 140 Mb/s if it is working properly.

If all the traffic is showing as going down one path then either ECP is mistaken or the hardware has a problem.

So we have some more questions ... ( I don't actually have a CI based machine here so I'm running from memory...sorry)

If a path has failed then in the error log and on the console, you should have seen messages like:
[...] change, Path #0. went from GOOD to BAD
( the paths on each card are variously referred to as 0 and 1 or A and B ).

Are you able to check for these messages like these?

Have you checked the amber error indicator LED for problems indicated on the actual cards?

$SHOW CLUS/CONT
should show both paths ( path A and path B ) on both cards as | A - B | if they are both active. If either letter is missing, the correspoonding path is inactive. Are you able to verify that VMS sees both paths as active ?


Hope this might lead us somewhere. Sorry to be asking so many questions.

Regards
Rob




Robert_Boyd
Respected Contributor

Re: Does HSJ Disks share CI paths (Two Paths)

When you run SHOW CLUSTER/CONTINUOUS you may want to beef up what you're showing in your display by adding a few things:

ADD CIRCUIT/ALL will display activities on the various adapters.

ADD PORT/ALL and ADD CONNECTION/ALL may provide additional interesting information.

Robert
Master you were right about 1 thing -- the negotiations were SHORT!
Shailender_1
Advisor

Re: Does HSJ Disks share CI paths (Two Paths)

Hi Robert,

Thanks for all your inputs, Don't feel sorry for anything. you are trying to help me with my query and to do so you need to understand the it thoroughly.

My both the Ci ciruit are good, Had checked before in CLUster and SDA utility I can see both the CI circuit as good, But my question here is how do I see both the ciruit are being working in Load sharng mode. How do I see Traffic counters on CI circuits to confirm both are taking part into disk data comunication.

I really appreciate all your dedication and support.

regards
Robert V Dixon
Occasional Advisor

Re: Does HSJ Disks share CI paths (Two Paths)

Without access to a CI based machine, I can't verify this, but I believe SDA should be able to give you what you want. Start by choosing a device at the end of the path and issue a
SDA> SHOW DEVICE
and work back thru to the controller.

e.g. SDA>SHOW DEVI DU

It will confirm (again) that both paths are good on PNA and PNB and it might show you traffic counters (or at least give a pointer to the counters).



Of course the performance monitoring tools should also be able to correctly report the traffic on the CI paths but I now suspect that the software has a bug and is incorrectly reporting all the traffic as being on the first path/cable.

Having checked the hardware indicator LED(s) and the view of the hardware from the VMS side, it looks to be the case that both CI cable sets are operating and I think we'd have seen error messages if either path had failed so I believe it is safe to assume that both cable sets are transferring data.

Since the performance monitor is reporting otherwise ( i,e, no traffic on a cable) , I would look to run some high traffic tranfer tests to show that the available bandwidth exceeds 9MB/sec on one CIPCA/star coupler.

Unless there really is a problem with the CI cabling, the CIPCA & start coupler should be able to move about 17.5 MB/sec ( about 140Mb/sec) but of course there may be other bottlenecks ( controllers, disks, ...)

Once you have traffic known to be exceeding 9MB/sec on the CIPCA , you can check the report of that traffic in the performance monitoring software and if it is in accurate you can report the evidence of the bug to the maintenance group.

Its probably worth reporting the bug ( one cable pair carrying 0 traffic) immediately since it is very unlikely that the current traffic is all down one cable pair per CIPCA or star coupler.