Sudden Slow File Access - Replica Problem?

[jaybo71]

HI,

We have mixed NW environment mostly NW 5, a couple of 4.11 servers still. 1 master replica, 2 r/w, 1 read only.. Tree has been pretty stable for years. All of a sudden file access times are bad to any server. Throughput is only 50 megs a minute as oppsed to 150+ under normal circumanstance.

Ran DSREPAIR on all servers (normal)
Checked replica Sync with DSTRACE (normal)
Checked out SYS vols on each machine (no probs)
Server utilizations are normal
Unloaded any 3rd party NLMs I could (Norton AV, etc)

Servers all look to be in decent shape hardware wise. Doesn't look like the switch is the culprit either (swapped out connections, reset cards)..

I have a sneaking suspicion that something is wrong with NDS itself but I can't prove it at this point. Any other way I can rule out a replica issue? Any other guidence would be sorely appreciated.

Thanks.

 

[marvhuffaker]

What changed, all of a sudden? I don't generally tie in file transfer performance with NDS.



Marvin Huffaker MCNE, CNE
Marvin Huffaker Consulting

http://www.redjuju.com

 

[terry712]

is it slow on all servers?

i wouldnt regard it a nds issue either if it's file access unless some change has cause it to tree walk or such like

is login etc ok

 

[jaybo71]

Thanks for the quick responses..

I made no significant Netware related changes prior to the slowdown. All servers are suffering similar sluggishness.

I want to blame the cenral switch but other non-netware servers are not experinecing any trouble.

Login speeds have not been affected..

Any idea where I should look at this point?

Thank you.

Jay

 

[igortsin]

I want to blame the cenral switch but other non-netware servers are not experinecing any trouble.
-------------------------------------
It still could be switch. I believe NW servers run on IPX protocol unlike the others.

 

[marvhuffaker]

Anything new or changed on the clients? And what are you using to gather your stats?

Marvin Huffaker MCNE, CNE
Marvin Huffaker Consulting

http://www.redjuju.com

 

[Sethos]

I suggest you take a LAN trace, then you will see in there what´s going on.

 

[TheLad]

Are there any errors on the server console(s)? Are your 4.11 servers sync'ing ok with the 5.1 servers?

-----------------------------------------------------
"It's true, its damn true!"
-----------------------------------------------------

 

[jaybo71]

I do not have a great way to gather statistics on this problem at the moment. I have been copying files from back and forth from multiple locations on site at off peak hours and comparing results. I have been getting some complaints from the user community about some slowness as well. Additionally the throughput on the backup jobs are also 1/2 or 1/3 as fast as jobs in weeks previous .

I have worked with our infrastructure guy looking at the switch. The connections look good at 100 Mb full duplex on each box.

No error messages on the console.

Other Windows and Unix boxes are plugged into the same switch but are functioning normally. I am wondering if there could be an IPX problem?

Thanks.

 

[marvhuffaker]

You might look in Monitor/ Lan/WAN Drivers -- NIC stats on your netware servers and see if any of the NICS are reporting an excessive number of errors.

Also if you have a bad NIC, it may be spewing out garbage and plugging up the bandwidth. A lan trace should help you find something like that.

Also verify that the NIC settings on the servers (speed/duplex) match that of the Switch ports. "Infrastructure Guy's" are notorious for changing settings without saying anything.

Marvin Huffaker MCNE, CNE
Marvin Huffaker Consulting

http://www.redjuju.com