FQuest Notify: SIX maintenance

[Terra]

04/15/99 at 1:00am EDT, I am planning to take the SIX server offline for 20 minutes or less...

I have been observing the server closely, and I'm and still not 100% confident in it's stability...

I am reverting back to original plans and swapping out the entire server with the new equipment, that we finished receiving today...

I do not forsee any difficulties with this as I will be moving the hard drives from one to the other and rebooting... This will also give me a chance to update the Linux kernel to 2.2.5 as well...

I was hoping that the new memory and power supply would have taken care of this, but I feel that the power problems may have potentialy stressed/damaged the motherboard...

Thank you for your patience with SIX... We are taking necessary measures to ensure that the hardware problems cease, and to **never** get caught up in this situation ever again...

--
Andrew Gillespie
Systems Administrator
FutureQuest.net

[Deb]

Running a bit late....

This will be closer to 2 am... and may be shaky for about 40 mins (but still goaling 20)

Deb

[flowersource]

It is after 4:00 AM and still down. Any estimate on when we will be up?

Mike

[Deb]

Seems it's back up (& down) now... I'm still waiting to hear from Andrew... I've a feeling he upgraded more then he was planning on... (I think he's in the final reboots now... but not positive)

Deb

[This message has been edited by Deb (edited 04-15-99).]

[Terra]

Tonight was smooth sailing, up until the last few minutes...

I have gone ahead and upgraded SIX to dual PII's with new Linux 2.2.5 SMP kernel build...

The delay was due to the SMP build, and a misconfigured setting that would not go away... I eventually had to nuke all the work I had done on the kernel and start over...

I believe that this one is going to work as desired...

One of those off-the-wall, twilight zone happenings... SIX is all brand new equipment and new kernel now

SIX seems to be responding quite well to the new configuration... I have already put her through some **serious** stress tests and she scaled very nicely with the processes and threads being split evenly between the cpu's...

--Andrew Gillespie
Systems Administrator
FutureQuest.net

[This message has been edited by Deb (edited 04-15-99).]

[frankc]

Andrew, just as a "heads up", FTP to SIX (hosannachurch.com) is extremely slow this morn, and it has been for about a week now since the last major rework of six. And it seems that pages are still being served slowly today and since the rework (altho we're on a LAN and I can't see a rate of transmission to give you hard proof).

Thanks for all your hard work on SIX! That one's been a real pain in the patoot for you (and he wonders as he types if 'patoot" is in the banned words list!).

------------------
Frank

Hosanna! Lutheran Church www.hosannachurch.com (http://www.hosannachurch.com)

[Terra]

With slowness, is it *only* when you connect to us, or intermittant throughout the day as you surf?

Network connectivity is the next item up for bids... http://www.aota.net/ubb/wink.gif

--
Terra
--I always wondered what the AutoBahn is like--
FutureQuest

[frankc]

It's only when connected to our site on SIX for web pages or FTP'ing. Everything else from everywhere else seems normal. Ran some tests and am getting consistent 54k throughput from CA, NY and FL sites with good latency. Tests were via
www.freetune.com/bt2kdemo/ (http://www.freetune.com/bt2kdemo/) this morn.

Took :28, :71, :40 and :78 to connect to our site, log in, go to passive mode and get a LIST of a subdirectory under cgi-bin with just one file in it. Used to be consistently well <:10 or so. Took :45 to transfer a 1954 byte ASCII file and :51 to get a LIST of that one original file after DELE the transferred file. Response to all commands seem very slow.

Also, it took :19 and :27 to check emails in two separate empty mailboxes.

Feel free to call me at 800-588-3209 and we can talk while running some tests.

Just ran Neotrace and to top it off, six nodes on alter.net and deltacom.net are balky with ping times <1400ms:

[Ucode]NeoTrace Version 1.22 - Registered
Destination: www.hosannachurch.com

-#--------------Node Name---------------IP Address--------RT*--High---Low---Avg-Tot---D-Who
1 frank.aquila.com 192.168.2.128 0 0 0 0 1 0 1
2 sc.aquila.com 204.95.88.209 2 2 1 2 17 0 1
3 s24.aquila.com 204.95.88.24 37 964 25 29 17 0 1
4 204.95.88.42 29 39 27 28 17 0 -
5 sl-gw15-chi-11-0-0-ts11.sprintlink.net 144.228.55.41 32 51 29 32 17 0 2
6 144.232.0.221 32 55 29 32 17 0 -
7 sl-bb11-chi-9-0.sprintlink.net 144.232.0.178 42 45 29 37 17 0 3
8 sl-bb3-chi-8-0-0.sprintlink.net 144.232.0.158 33 67 31 33 17 0 3
9 0.0.0.0 - 1567 1567 522 18 17 -
10 101.atm3-0.xr2.chi4.alter.net 146.188.208.6 38 1407 33 36 17 0 4
11 294.atm2-0.tr2.chi4.alter.net 146.188.208.234 1471 1878 1174 1373 17 1 4
12 106.atm7-0.tr2.atl1.alter.net 146.188.136.9 1033 1975 1033 1263 17 3 4
13 198.atm7-0.xr2.atl1.alter.net 146.188.232.101 1320 1732 1196 1294 17 1 4
14 194.atm9-0-0.gw1.atl1.alter.net 146.188.232.45 1474 1862 1205 1354 17 0 4
15 itcdeltacom-gw.customer.alter.net 157.130.65.82 1556 1782 1254 1501 17 1 4
16 mercury-gw.fr.customer.deltacom.net 209.192.0.34 1576 1852 1307 1642 16 3 5
17 www.hosannachurch.com 209.192.41.27 1472 2252 1334 1484 17 5 6
-----------------------------------------------------------------------------------------------
*All times in milliseconds (ms), D=Dropped packets
[/quote]

THANKS!

[This message has been edited by frankc (edited 04-15-99).]

[frankc]

Andrew: YAY--it's not SIX!

With most of the trouble being network-oriented, is there a site or newsgroup with real current info on network problems? I'm thinking something other than Internet Traffic Report, altho that's very helpful (right now, the US router index is 36 and falling; all are yellow and red except for Toronto!).

As always, thanks for all your hard work 'n stuff!

[Terra]

I believe that UUnet/Alter.net is having some growing pains themselves...

I do know that they are upgrading some major pipes throughout the US and installing/upgrading to OC-192 lines...

Many of the glitches we are seeing is a direct result of these upgrades...

It's sorta the same boat we are in with all of our upgrades... There are some very rough spots along the way, but in the end the final result is 10 times better than it was before...

I remember two months ago, everyone wanted off of SprintLink, due to it's severe problems and downtime... After SprintLink sorted out the situation, TAZ has been running solid as a rock with great connection times...

I still don't like SprintLink, and probably never will, but the problematic backbone right now is UUnet due to all the upgrades they are doing...

The Internet is far too dynamic to even begin to comprehend, and how it even manages to run with all the network load and peoples expectations, is *way* beyond me... It's nothing short of amazing...

--
Terra
--Time and Patience for a better tomorrow--
FutureQuest