Minutes Waterloo Polaris Advisory Group June 21, 2000

Attendees:

Daniel Delattre AHS Trevor Bain Environmental Studies
Nevil Bromley Arts Tim Farrell IST
Erick Engelke Engineering Computing Stephen Sempson Science
Hon Tam Engineering Computing  

Submitted items:

Q: (Nevil) Update to Winzip to suport ARC files.
A: (Hon) Winzip does support with the capability of File Type. If you try, File, Open Archive, a window will show with File Name and File Type at the bottom. Choosing all archives with allow one to pick up on the ARC file types. Another way is that Winzip will accept a string for another program if it does not support the file type. This can be found in Options, Program Locations. In the middle of this window, it does state that some non-Zip formats require external programs. Three listed are ARJ, LHA, and ARC.
Q: (Nevil) Still getting sysctl's in the middle of the day (Duane, Vic and Carl were all present when it started).
A: (Hon) Does this happen all of the time.

(Nevil) Yes, and in the labs too. Would you have any idea why this is happening?
(Erick) No. It should be random between 1 and 5 in the morning.

Information items:

(Nevil) This is a request for Ray. Carl H. would like to have another Netscape Plug-in installed on the network. We also have a software request which was just received yesterday.

(Erick) We will gladly give permissions on the files if you would like.
(Nevil) Should I talk to Paul McKone about this/
(Erick) Yes, I am not sure if this got moved to the other Engineering machines.

(Tim) Should all files on the AppFiler's be encripted?

(Erick) The trick we found does not work with all files. It does not hurt to try, you cannot damage anything. Will attempt in the future to look at improving this.
(Tim) Roger Watt is considering the ramifications.
(Nevil) I have seen people using the AppFiler's from home to run software.

(Nevil) It looks like Engineering has two AppFilers.

(Erick) Yes, one is a backup.
(Nevil) Arts is looking at having a second one for speed in our network.
(Erick) Has IST improved the 5516 there?
(Nevil) Yes, and we have noticed no discernable difference.
(Steph) In Science we have been told that there is network problems here. In analyzing the situation, we have had Gary Cairney check the status of our switchs which run between Scifile and our NFS mounted host to the Hoover services. Our traffic shows no errors. I am thinking that the backbone switch connections are the problem. Do you use Bruce's RWTest program still. I am thinking of running the program internally in Science, and then going to Engineering to see the response difference.
(Erick) Yes, we do use it and it would be a good test.

From 129.97.80.29 to Engfile
13:29:31 S:\>rwtest /init
8191ializing 8192K...0
13:29:54 S:\>rwtest
Shows V=version R=read_percent D=home-dir K=K-per-sec
A=average_response_in_milliseconds S=std_response M=max_response
N=number_of_ops_which_exceed_max/2 C=close/opens L=reps
Run "rwtest /help" for options

Performing test...

[RWTEST] V=8 R=100 D=S:\ K=951 A=8.23 S=4.96 M=34.00 N=0/1288@0.00 C=4 L=0
[RWTEST] V=8 R=100 D=S:\ K=910 A=8.53 S=6.63 M=132.00 N=47/1215@17.00 C=2 L=1
[RWTEST] V=8 R=0 D=S:\ K=704 A=12.32 S=13.55 M=240.00 N=4/891@66.00 C=0 L=2
[RWTEST] V=8 R=0 D=S:\ K=732 A=11.39 S=20.94 M=602.00 N=3/969@120.00 C=0 L=3
[RWTEST] V=8 R=99 D=S:\ K=926 A=8.95 S=6.82 M=152.00 N=0/1206@301.00 C=3 L=4
[RWTEST] V=8 R=100 D=S:\ K=975 A=8.49 S=5.17 M=46.00 N=0/1270@76.00 C=3 L=5
[RWTEST] V=8 R=0 D=S:\ K=665 A=12.19 S=14.07 M=306.00 N=39/901@23.00 C=0 L=6
[RWTEST] V=8 R=0 D=S:\ K=742 A=11.44 S=11.61 M=251.00 N=1/964@153.00 C=0 L=7
[RWTEST] V=8 R=79 D=S:\ K=853 A=9.73 S=17.53 M=551.00 N=1/1124@125.50 C=1 L=8
[RWTEST] V=8 R=81 D=S:\ K=811 A=10.01 S=23.27 M=532.00 N=3/1087@275.50 C=2 L=9
[RWTEST] V=8 R=80 D=S:\ K=872 A=9.40 S=8.89 M=210.00 N=0/1162@266.00 C=2 L=10
[RWTEST] V=8 R=78 D=S:\ K=807 A=10.15 S=26.17 M=653.00 N=2/1073@105.00 C=2 L=11
[RWTEST] V=8 R=80 D=S:\ K=765 A=10.91 S=32.27 M=677.00 N=3/1004@326.50 C=1 L=12
[RWTEST] V=8 R=80 D=S:\ K=869 A=9.21 S=17.81 M=583.00 N=1/1177@338.50 C=2 L=13

(Erick) Bruce fixed some of our problems by taking out the IST switch and putting in one of Engineering's
(Hon) See about Roger putting a packet sniffer in the Physics building. I think the problems are stemming form there.
(Tim) We had a new switch installed for use with all of our servers and problems too.

(Daniel) SPSS 10 - crashes with multiple users, has anyone noticed similar problems or does anyone know how to fix this?
(Daniel) That was an old question we had. This could have been the new switches which were installed at that time. SPSS is working fine now.
(Hon) Watch for 100Mb Full Duplexing, there seem to be problems with Samba mounts. With the 2916's and 1924's, test copying back and forth caused long times for the copy to finish. If set to 100 Half the copying only goes on for a couple of seconds.
(Tim) What type of Ethernet card are you using?
(Hon) Dlink 530's and Intel 100+ Management Card. The Intel is more expensive, but a better card with the 82559 chip set.

(Hon) I was wondering if Science is thinking of joining soon with the Win2K Domain.
(Steph) Yes, we have a server now, not the best, but we are ready for trials.
(Hon) I am working out a few problems that have shown up in Arts, but after will look into Science.
(Nevil) What do you recommend for this type of server?
(Erick) What ever you can afford.
(Hon)
 

Polaris domain (for production):
Eng2K-1 (main DC, IIS and GC owner)
P3-500 (Katmai core, 512KB L2 cache)
Asus P2B-DS M/B
256MB PC100 SDRAM
2 x 18G SCSI-3 (LVD) 10,000RPM IBM HDD
Intel Pro100+ Management NIC
SCSI-2U CDROM
Eng2K-2 (another DC and backup IIS)
P3-500 (Katmai core)
Asus P2B-F M/B
256MB PC100 SDRAM
2 x 18G SCSI-3 (LVD) 10,000RPM IBM HDD
Intel Pro100+ Management NIC
IDE CDROM
Clipper (another DC, RIS, DHCP, etc)
2 x P3-650 (Cu-mine core, 256KB L2 cache)
Asus P2B-DS M/B
512MB PC100 SDRAM
36G SCSI-3 (LVD) 10,000RPM IBM HDD
Intel Pro100+ Management NIC
SCSI-2U CDROM
Engtest domain (for development):
Kasumi-chan (main DC, GC owner, RIS, DHCP, etc.)
2 x P2-350 (Deschutes core, 512KB L2 cache)
Asus P2B-DS M/B
396MB PC100 SDRAM
36G SCSI-3 (LVD) 10,000RPM IBM HDD
Intel Pro100+ Management NIC
SCSI-2U CDROM


Created by: sempson@sciborg.uwaterloo.ca 2000/06/22
Revised by: sempson@sciborg.uwaterloo.ca 2000/06/22
Revised by: tim@ist.uwaterloo.ca 2000/10/17