40m QIL Cryo_Lab CTN SUS_Lab TCS_Lab OMC_Lab CRIME_Lab FEA ENG_Labs OptContFac Mariner WBEEShop
  40m Log, Page 79 of 339  Not logged in ELOG logo
ID Date Author Type Categoryup Subject
  401   Tue Mar 25 13:21:25 2008 AndreyUpdateComputersc1susvme2 is not behaving itself again
  403   Tue Mar 25 16:34:47 2008 robUpdateComputersc1susvme2

Quote:

Quote:
c1susvme2 isn't behaving itself. It keeps getting out of sync and/or giving a red status light.

After going through the usual restart procedures a few times (unsuccessfully) we power cycled the c1susvme & c1sosvme crates. We think everything came back okay.

We still can't get the status and CRC (cyclic redundancy check) to return to normal on c1susvme2. If Alex is around tomorrow please ask him to take a look.


I rebooted it again this morning. The ASS machine is currently not running its process, for whatever reason (someone turn it off?). Let's leave it like this for a day and see how the c1susvme2 does. The other recent change is Steve's install of a cooling fan--maybe that's causing the problem.


Now c1susvme1 is joining the action. Since leaving the ASS off doesn't change anything, we can probably absolve it of blame. I now suspect the 4-pin LEMO cables going from the CLK DRIVER modules to the clock fanout modules. These cables are being squeezed/shaken by Steve's new fan setup, and may have been the culprit all along. John will do some testing to see if they are indeed the problem.
  405   Wed Mar 26 22:26:15 2008 JohnUpdateComputersc1susvme
I removed the fan and tweaked the timing cables to see if they were the source of our problems. I saw no effect. I'm leaving the fan off for the moment to see if that helps. It is on top of the filing cabinet next to my desk.
  406   Fri Mar 28 16:18:18 2008 robUpdateComputersc1susvme2 status
c1susvme2 is getting worse and worse. it won't run for more than ~45 minutes without fatally de-syncing. for now I've turned off c1iovme (which sends the MCL signal) to see if that's causing the problem. next I'll swap the boards for c1susvme1 and c1susvme2 to see if it's the cpu (or maybe the RFM card) itself, rather than the timing/pentek systems.
  408   Mon Mar 31 14:14:16 2008 robUpdateComputersc1susvme2 status

Quote:
c1susvme2 is getting worse and worse. it won't run for more than ~45 minutes without fatally de-syncing. for now I've turned off c1iovme (which sends the MCL signal) to see if that's causing the problem. next I'll swap the boards for c1susvme1 and c1susvme2 to see if it's the cpu (or maybe the RFM card) itself, rather than the timing/pentek systems.


I swapped the processors for c1susvme1 and c1susvme2. So for now, to startup, you should ssh into c1susvme1 and run the startup.cmd for c1susvme2, and vice versa.
  412   Thu Apr 3 18:46:04 2008 AndreyConfigurationComputers"Network switch board" and "c1pem1 crate" were touched

While working with the weather station, I did two things that potentially (with a very small probability) might influence the smooth work of other processors/computers.
I did the following on Wednesday, April 2nd, in times between 1PM and 3PM.

(1) I turned off for several seconds and returned into the initial position the switch-key on the rack with computer (processor) 'c1pem1' in order to reboot processor 'c1pem1'. The turning off/on of that key-switch was repeated several times.

(2) I pulled gently the whole "Network-Switch Board" towards me in order to replace an ethernet cat 5 cable going into the board form the processor 'c1pem1'. Some other connections of other ethernet cables might be flimsy, and then other people in 40-meter might have problems with computers other than 'c1pem1'. It should not happen, but in case of extraordinary behaviour of any other computer in our lab, people should check the connectors on the network-switch board. It is located near the middle of Y-arm. See picture.
Attachment 1: Computer_Rack.JPG
Computer_Rack.JPG
  416   Mon Apr 7 16:42:56 2008 ranaUpdateComputerseLog intermittent
Phil Ehrens restarted Dziban again this morning. Looks like its still crashing each Monday around 8 AM.

Here is the latest suspect:

http://open.itworld.com/5040/reboot-unix-nlsunix-071225/page_1.html
  419   Tue Apr 15 18:44:25 2008 ranaConfigurationComputersRosalba
There is a new computer in the control room -- its called Rosalba,
in keeping with our naming convention. Its a quad-core machine that
Dmass found for cheap somewhere; we've installed the CentOS on it
that Alex recommended.

Its a 64-bit Linux and so that's going to cause some problems. Alex has done this
before and so we have some confidence that we can get our regular tools (DTT, Dataviewer)
to run on it.

I have made a new apps tree for all of our future 64-bit Linux machines. So far, there is
a 64-bit firefox and a 64-bit matlab in there. As we start using this machine some more, we
will be forced to install more 64-bit Linux stuff.

We also didn't have enough network cables to run to both linux3 and rosalba. Andrey has decided that we
should not ditch linux3 and so he will run another cable for it tomorrow.
  421   Wed Apr 16 10:20:01 2008 AndreyUpdateComputersRosalba and linux3

Quote:
There is a new computer in the control room -- its called Rosalba,
in keeping with our naming convention. Its a quad-core machine that
Dmass found for cheap somewhere; we've installed the CentOS on it
that Alex recommended.

Its a 64-bit Linux and so that may cause some problems. Alex has done this
before and so we have some confidence that we can get our regular tools (DTT, Dataviewer)
to run on it.

I have made a new apps tree for all of our future 64-bit Linux machines. So far, there is
a 64-bit firefox and a 64-bit matlab in there. As we start using this machine some more, we
will be forced to install more 64-bit Linux stuff.

We also didn't have enough network cables to run to both linux3 and rosalba. Andrey has decided that we
should not ditch linux3 and so he will run another cable for it tomorrow.


The ethernet cable for linux3 was installed on Wednesday morning. Now linux3 has Internet connection again.
  444   Thu Apr 24 22:06:47 2008 AndreySummaryComputersEthernet Cables and Hubs
Today in the morning (between 8.30AM and noon) Joe and I were working on understanding which ethernet cables connect "processors controlling the work of equipment in the interferometer room" and "Internet hub in the computer room".

Firstly, we took off several times the blue ethernet cables from the router located near ETMX in the morning. We were trying to understand which port in the hub is responsible for the interaction with that processor.

Secondly, we were working on reviving the connection with the computer controlling vacuum in the interferometer.

Later in the middle of the day (around 2PM) Joe continued some work with ethernet cables without me. We plan on continuing the cable work on Friday morning. A better and more detailed elog will appear then.
  447   Fri Apr 25 11:33:40 2008 AndreyConfigurationComputersComputer controlling vaccum equipment

Old computer (located in the south-end end of the interferometer room) that was almost unable to fulfill his duties of controlling vacuum equipment has been replaced to "Linux-3". MEDM runs on "Linux-3".

We checked later that day together with Steve Vass that vacuum equipment (like vacuum valves) can be really controlled from the MEDM-screen 'VacControl.adl'.

Unused flat LCD monitor, keyboard and mouse (parts of the former LINUX-3 computer) were put on the second shelf of the computer rack in the computer room near the HP printer.
  449   Fri Apr 25 13:53:11 2008 josephbSummaryComputersNetwork setup
This is the promised more in detail summary from Andrey's log ID 444.

What we did was go around to each hub, one at a time, unplug the network connection, and figure out which light on which hub went out. We then, went back to the control room, confirmed that we were still able to talk to the devices connected to the hub, and if not, rebooted them. This process was repeated for each hub.

As it stands, the hubs located at the ends of arms (in racks 1X4 and 1Y9) are connected to the really old 24 port 10 Base T hub located in 1Y7. In addition, the 5 port SMC hub is plugged into the 8 port SMC switch in 1Y5 (which actually has enough ports to simply move all the connections over to it, so I'm not sure why there are two...).

All other hubs/switches are connected back to the control room 24 port switch.

Attached is a simple diagram of the network connections for the 40m lab.
Attachment 1: 40m_network_90.pdf
40m_network_90.pdf
  453   Sat Apr 26 11:21:15 2008 ajwOmnistructureComputersbackup of /cvs/cds restarted
The backup of /cvs/cds (which runs as a cron job on fb40m; see /cvs/cds/caltech/scripts/backup/000README.txt) 
has been down since fb40m was rebooted on March 3.
I was unable to start it because of conflicting ssh keys in /home/controls/.ssh .
With help from Dan Kozak, we got it to work with both sets of keys
( id_rsa, which allows one to ssh between computers in our 113 network without typing a password,
 and backup2PB which allows the cron job to push the backup files to the archive in Powell-Booth).

It still goes down every time one reboots fb40m, and I don't have a solution.
A simple solution is for the script to send an email whenever it can't connect via ssh keys
(requiring a restart of ssh-agent with a passphrase), but email doesn't seem to work on fb40m.
I'll see if I can get help on how to have sendmail run on fb40m.
  456   Sun Apr 27 18:11:58 2008 robDAQComputersbr40m?

The testpoint manager (which runs on fb40m) crashed this afternoon. Upon re-starting it, I found there was a rogue dtt process on op440m and also a daqd daemon running on br40m. One or both of these caused the tpman to crash. br40m is the frame broadcaster, which is never used here as we don't run DMT. I killed the daqd process there.

The way to find if there is a rogue process is to watch the output to the console from the tpman when you start it:

Allocate new TP handle 56 by 131.215.113.203
Allocate new TP handle 57 by 131.215.113.203
Allocate new TP handle 58 by 131.215.113.203
Allocate new TP handle 59 by 131.215.113.203
Allocate new TP handle 60 by 131.215.113.203
Allocate new TP handle 61 by 131.215.113.203
Allocate new TP handle 62 by 131.215.113.203
Allocate new TP handle 63 by 131.215.113.203
Allocate new TP handle 64 by 131.215.113.203
Allocate new TP handle 65 by 131.215.113.203
Allocate new TP handle 66 by 131.215.113.203
Allocate new TP handle 67 by 131.215.113.203
Allocate new TP handle 68 by 131.215.113.203


If you see something like this, with a new TP handle being allocated every few seconds, you need to log in to the corresponding host and kill whatever process has run away.
  457   Sun Apr 27 22:57:15 2008 ajwDAQComputersbr40m?

Quote:

The testpoint manager (which runs on fb40m) crashed this afternoon. Upon re-starting it, I found there was a rogue dtt process on op440m and also a daqd daemon running on br40m. One or both of these caused the tpman to crash. br40m is the frame broadcaster, which is never used here as we don't run DMT. I killed the daqd process there.

The way to find if there is a rogue process is to watch the output to the console from the tpman when you start it:

Allocate new TP handle 56 by 131.215.113.203
Allocate new TP handle 57 by 131.215.113.203
Allocate new TP handle 58 by 131.215.113.203
Allocate new TP handle 59 by 131.215.113.203
Allocate new TP handle 60 by 131.215.113.203
Allocate new TP handle 61 by 131.215.113.203
Allocate new TP handle 62 by 131.215.113.203
Allocate new TP handle 63 by 131.215.113.203
Allocate new TP handle 64 by 131.215.113.203
Allocate new TP handle 65 by 131.215.113.203
Allocate new TP handle 66 by 131.215.113.203
Allocate new TP handle 67 by 131.215.113.203
Allocate new TP handle 68 by 131.215.113.203


If you see something like this, with a new TP handle being allocated every few seconds, you need to log in to the corresponding host and kill whatever process has run away.


I *think* Alex is responsible for the daqd daemon running on br40m (he set up some new stuff recently, a data concentrator and broadcaster); I'll make sure he sees this post.
  463   Thu May 1 12:46:02 2008 josephbConfigurationComputersNodus gateway is up
The computer Nodus is now acting as a gateway machine between the GC network and the martian network in the 40m. It has the same passwords as the rana gateway machine.

Its name on the GC side is nodus (ip: 131.215.115.52) and on the martian side is nodus113 (ip: 131.215.113.200). Will need to update the hosts file on the control room machines so you can just use the name nodus113 rather than the full ip.

Software is still being added to the computer, and it will remain in parallel with the rana gateway machine until everything has been working properly for a week or so.
  464   Mon May 5 11:04:30 2008 robOmnistructureComputersNetwork setup

Mafalda was not connected to the network, and so our DMF-based seisBLRMS has not been running for ~1 week. I traced this to a broken ethernet cable connecting mafalda to the network switch in the rack next to the B&W printer. This cable has a broken connector at the switch side, which means it can't stay connected if there's any tension. It needs to be replaced.
  473   Fri May 9 10:15:36 2008 josephbUpdateComputersNodus has moved
Steve and myself moved Nodus from under the table in the control room, to just above the Rana computer in the control room rack.
  475   Tue May 13 10:38:28 2008 steveUpdateComputersrfm network is down
The RFM network went down yesterday around 5pm
Only c1susvme1 is alive but it's timing is off.
Andrey is bringing the network up.

Andrey wants to make an addition first that our situation is very much similar to that described by Rana in his elog entry # 353 (March 03). All of the rectangular boxes are red, except for SUS1-c1susvme1 and AWG (only these two rectangles are green).
  476   Wed May 14 13:14:19 2008 AndreySummaryComputersReflective Memory Network is restored

Reflective Memory Network is restored, all watchdogs and oplevs are returned to the "enabled" state.

In order to revive the computers, several things were done.

1) Following Mr. Adhikari's elog entry #353, I walked around the interferometer room, and switched off the power keys in all crates with computers whose names are contained in the MEDM Reflective Memory screen, including the rack with the framebuilder. By the way, it was nontrivial to find the switch in the 1Y4 crate that would shut off/on processors "c1susvme1" and "c1susvme2": the switch turned out to be located at the rear side of the crate, and it is not a key but it is a button.

2) I was trying to follow wiki-40 computer restart procedures, but every time that I was trying to run "startup.cmd" screen from the corresponding target subdirectory, I got the error message "Device or resource busy".
By the way, one more thing was learned: if you firstly open in terminal burtgooey, select the snap file, then reboot the processor, and then will try to burt-restore it, you will get the message "Status Not OK". In order to really burt-restore the processor which was recently rebooted, you need to close the terminal with burtgooey and open burtgooey in a new terminal window which should be opened after rebooting the processor.

Feeling that my activities according to wiki-40 procedures do not revive computers, I invited Alex Ivanov.

3) Alex tried to touch the memory card in "c1iovme" in rack 1Y2, because once before this card failed causing network problems, but this did not help.

4) We shutted off and restarted again (pressing the power-switching button) the black Linux machine "c1dcuepics" (located in the very bottom below the framebuilder). Alex says that this machine is responsible for all EPICS. It was not restarted for 182 days, and probably some process there went wrong.

After restarting this machine "c1dcuepics" we were able to follow wiki-40 procedures for restarting all other computers (whose names are on the MEDM RFM network). We ran correcponding "startup.cmd" files and burt-restored them without error messages.

Now all the computers work and communicate in a proper way.

Mr. Joseph Betzwiezer was helping me with all these activities (we decided that it is more important that cameras for now), thanks to him. But our joint skills turned out to be insufficient, so Alex Ivanov's contribution was the most important.
  477   Wed May 14 14:05:40 2008 AndreyUpdateComputersComputer Linux-2, MEDM screen "Watchdogs"

Computer "Linux-2", MEDM screen "C1SUS_Watchdogs.adl": there is no indication for ETMY watchdogs, everything is white. There is information on that screen "C1SUS_Watchdogs.adl" about all other systems (MC, ETMX,...), but something is wrong with indicators for ETMY on that particular control computer.
  486   Sun May 18 18:59:15 2008 ranaConfigurationComputerscron and hosts
I added rosalba to the hosts file for the control room machines (131.215.113.103).

I also removed the updateddb cron from our op440m crontab because it was running at 5 PM
even though I had set it to run at 5:57 AM. If it still runs then, it must be because of
another crontab.
  492   Thu May 22 11:25:19 2008 josephbConfigurationComputers 
One of the new Netgear Prosafe 24 port switches was mounted in the 1X4 rack,, roughly in the middle, away from the top and bottom rack mounted electronics. At the moment, its IP has been set to 131.215.113.250, gateway 131.215.113.2 (which is what I saw as the only listed gateway on linux1 using route -n) and mask 255.255.255.0.

I'm planning to set the next three IP address for the switches as *.251, *.252 and *.253, which don't look to have been used yet.
  495   Sun May 25 16:20:27 2008 ranaConfigurationComputersjoinPDF
I have installed joinPDF 2.1 on rosalba. Since its written in Java, I didn't have to tinker with it at all to work on a 64-bit machine. Now Caryn can put all of her plots into 1 file.
  498   Sun May 25 21:14:14 2008 tobinConfigurationComputersEPICS proxy server
I set up an EPICS gateway server on Nodus so that we can look at 40m MEDM screens from off-site.
The gateway is set up to allow read access to all channels and write access to none of them.

The executable is /cvs/cds/epics/extensions/gateway; it was already installed. A script to start
up the gateway is in target/epics-gateway. For the time being, I haven't set it up to start itself
on boot or anything like that.

To make it work, you have to set the environment variable EPICS_CA_ADDR_LIST to the IP address of
Nodus. For instance, something like this should work:

setenv EPICS_CA_ADDR_LIST 131.215.115.52

On Windows you can set up environment variables in the "System" Control Panel. On one of the tabs
there's a button that lets you set up environment variables that will be visible to all programs.

On Andrey's machine I installed the Windows EPICS extensions, i.e. MEDM and its friends. I also
installed the cool Tortoise SVN client which lets you interact with SVN repositories through
the windows explorer shell. (The right-click menu now contains SVN options.) I checked out
the MEDM directory from the 40m SVN onto the desktop. You should be able to just right-click in
that window and choose "SVN Update" to get all the newest screens that have been contributed to
SVN; however, there are currently some problems with the 40m SVN that make that not go smoothly.

At the moment on Andrey's (Windows) machine you can go into the MEDM folder and double-click on
any screen and it will just work, with the exception that not all the screens are installed
due to SVN difficulties.
  501   Wed May 28 12:51:32 2008 josephbConfigurationComputersTwo more switches mounted
Two more Prosafe 24 port switches have been mounted in the racks, one in 1Y9 and one in 1Y6. (The first one was placed in 1X4).

The one in 1Y9 has been set to an IP address of 131.215.113.251, while the one in 1Y6 is set to 131.215.113.252, and these have been labeled as such.
  508   Fri May 30 21:30:15 2008 tobinConfigurationComputerssvn on solaris
I installed svn on op440m.  This involved installing the following packages from sunfreeware:

apache-2.2.6-sol9-sparc-local  libiconv-1.11-sol9-sparc-local   subversion-1.4.5-sol9-sparc-local
db-4.2.52.NC-sol9-sparc-local  libxml2-2.6.31-sol9-sparc-local  swig-1.3.29-sol9-sparc-local
expat-2.0.1-sol9-sparc-local   neon-0.25.5-sol9-sparc-local     zlib-1.2.3-sol9-sparc-local
gdbm-1.8.3-sol9-sparc-local    openssl-0.9.8g-sol9-sparc-local

The packages are located in /cvs/cds/caltech/apps/solaris/packages.  The command line to install
a package is "pkgadd -d " followed by the package name.  This can be repeated on nodus to get
svn over there.  (Kind of egregious to require an apache installation for the svn _client_, I 
know.)
  509   Sun Jun 1 19:25:10 2008 ranaConfigurationComputersnew monitor on op440m
I installed the new 24" flat screen on op440m. I increased the screen resolution from 1280x1024 to 1900x1200 using
the obscure 'fbconfig' command. You can type Google it if you want.

The old monitor is on the surplus cart. If you are reading this and think you might walk from the 40 over to
Bridge, please wheel the cart full of old computer equipment (on the north side of the control room) over to Larry.

I also copied over all the images on the D40 to a folder on Kirk's computer and deleted the originals.

Dan Busby also visited us last week to help us move the drill press from the Y arm down into the sub basement
of W Bridge.
Attachment 1: Andrey-440.jpg
Andrey-440.jpg
Attachment 2: Busby08.jpg
Busby08.jpg
  510   Sun Jun 1 19:39:35 2008 tobinConfigurationComputerselog, etc
Phil Ehrens gave me a DVD of the 40m elog, apache, and (Jamie's) SVN archive.
I copied it to nodus:/home/controls/dvd-from-ehrens.  Once we get the elog
running on nodus, we can copy the datafile over again from dziban (so that
we don't lose any elog entries) and switch over.
  513   Tue Jun 3 10:19:45 2008 tobinConfigurationComputersbig machine
Several of us transported the big new awesome Sun box from Bridge over to
the 40m last week. If I recall correctly, it's a SunFire X4600 with
something like sixteen 64-bit AMD processor cores at 2.8 GHz. It sounds
like a jet engine when it starts up (before the cooling fans are throttled
back) and has four power supplies (each with its own connection
to the wall). It has slick removable hard disks and fan units too. Our
working name for it is "megatron".

Anyway. It came with two hard disks, one with Solaris 10 installed. I took
the other hard disk over to Alex, who copied a Realtime Linux installation
onto it. Alex says it boots and runs fine.

It remains for you guys to install the machine onto rails and install the
whole thing into a rack. Before it goes into service as a realtime control
machine, you might as well install Matlab on it and do some heavy-duty
computation.

  514   Tue Jun 3 10:40:27 2008 tobinConfigurationComputersnew dataviewer
Alex let me know the secret location of the latest dataviewer executable for Linux. It is:

http://www.ligo.caltech.edu/~aivanov/upload/dv/Control/dc3

If your linux dataviewer on linux2 has the "year field not filled in" bug, you should download this into /usr/local/bin/dc3 (after making a backup of that file).

It looks like there's no dataviewer installed on rosalba yet. We should figure out a better directory layout for the linux machines; currently dataviewer is installed locally on linux2. It should be in /cvs/cds/caltech/apps/linux/something so that all the linux machines see the same installation.
  538   Wed Jun 18 16:07:57 2008 robSummaryComputersRFM network down

The RFM network tripped off around noon today. It's still down. The problem appears to be with the EPICS interface (c1dcuepics). Trying to restart one of the end stations yields the error: No response from EPICS.

Possible causes include (but not limited to): busted RFM card on c1dcuepics, busted PMC bus on c1dcuepics, busted fiber from c1dcuepics to the RFM switch. We need Alex.
  544   Wed Jun 18 18:50:09 2008 ranaUpdateComputersIt can only be attributable to human error. (HAL - 2001)
There has been another one of "those" events and all of the front end machines are down.

We poked around and Rob determined that the FEs can't get the EPICS data from EPICS. The
dcuepics machine is hooked up and running and all of the epics binaries are running. We also
tried resetting its RFM switch as well as power cycling the box using the "poweroff" command.


Not a sausage.

Rob points out that although the Signal Detect lights are on on the cards, the 'Own Data' light
is not on on the dcuepics' card although it is on for some of the cards on the other boxes.


We have placed messages with the Russian. If anyone sees him, don't let him go without fixing things.
Also, make sure to follow him around with notepad and possibly a camera to record what it is that
he does. If he's muttering, maybe try to use a sensitive hidden sound recorder.
  545   Thu Jun 19 15:52:06 2008 AlbertoConfigurationComputersMeasure of the current absorbed by the new Megatron Computer
Together with Rich Abbot, sam Abbot and I measured the current absorbed by the new Megatron computer that we installed yesterday in the 1Y3 rack. The computer alone absorbs 8.1A at the startup and then goes down to 5.9A at regime. The rest of the rack took 5.2A without the computer so the all rack needs 13.3 at the startup and the 11.1A.

We also measured the current for the 1Y6 rack where an other similar Sun machine has been installed as temporary frame builder and we get 6.5A.


Alberto, Rich and Sam Abbot
  586   Fri Jun 27 19:59:44 2008 JohnUpdateComputersc1iovme
C1susvme2 and C1iovme crashed which sent the optics swinging and tripped the watchdogs.

Koji and I were able to restore c1susvme2 without any trouble.

We have been unable to revive c1iovme. We have tried telneting in and running startup.cmd,
the process runs for a while then hangs with "DAQ init failed -- exiting".

Resetting the board doesn't help. I didn't try keying the whole crate.

All optics are back to normal with damping restored.
  587   Sat Jun 28 03:10:25 2008 robUpdateComputersc1iovme

Quote:
C1susvme2 and C1iovme crashed which sent the optics swinging and tripped the watchdogs.

Koji and I were able to restore c1susvme2 without any trouble.

We have been unable to revive c1iovme. We have tried telneting in and running startup.cmd,
the process runs for a while then hangs with "DAQ init failed -- exiting".

Resetting the board doesn't help. I didn't try keying the whole crate.

All optics are back to normal with damping restored.


I tried keying the crate, then keying the DAQ controller & AWG, then powering down & restarting the framebuilder.
On coming up, the framebuild doesn't start a daqd process, and I can't get one to start by hand (it just prints "652", and then stops).
No error messages and daqd doesn't appear in the prstat.

I then tried keying the DAQ controller again (after the fb0 reboot), which blew the watchdogs on all the suspensions. So then I went around and keyed all the crates.

Now, the suspension controllers are back online. Still no c1iovme, and now the framebuilder/DAQ/AWG are also hosed. We can try keying all the crates again, in the order that Yoichi did last week.

After some more poking around, I found the daqd log file. It's now complaining about

Jun 28 03:00:39 fb daqd[546]: [ID 355684 user.info] Fatal error: channel `C1: PSL-FSS_MIXERM_F' is duplicated 126

This is the second error message like this. It first complained about C1: PSL-FSS_FAST_F, so I commented that out of C1IOOF.ini and rebooted the framebuilder (note this is an actual reboot of the full solaris machine). Eventually I discovered that C1IOOF.ini and C1IOO.ini are essentially identical. They presumably will keep getting these duplicate channel errors until one of them is completely removed.

C1IOO.ini has a modification time of seven PM on Friday night. Who did this and didn't elog it? I've now modified C1IOOF.ini, and I don't remember when it was last modified.
  588   Sat Jun 28 14:56:44 2008 JohnUpdateComputersini files
In short, I was editing the ini files yesterday evening, I didn't e-log it and after some investigation this afternoon it apears
that I am to blame for all the computer problems which followed.


I wanted to edit C0EDCU.ini and C1IOOF.ini to change C1: PSL-FSS_FAST to a fast channel as C1: PSL-FSS_FAST_F
was dead.

I opened these files and made backups. It appears this is where it all went awry. My backup for C1IOOF
is called C1IOO.ini.090627 i.e. missing the F.

Later c1susvme2 and c1iovme crashed. After failing to bring c1iovme back I wondered if my edits had
caused the problems so I restored the back up files. It appears that here I wrote over C1IOO with my backup of
C1IOOF (presumably because I had made a typo in the name).

To remedy the situation we could restore C1IOO from e.g. chans/archive/C1IOO_080618_160028.ini

No excuses for not e-logging this activity.
  589   Sat Jun 28 23:23:50 2008 JohnUpdateComputersRebooting
All of the computers are now showing green lights.

Remaining problems:

Alignment scripts are failing with "ERROR: LDS - NDS server error #13"
I think this is a server transmission error.

Dataviwer shows all channels as zero.
  592   Sun Jun 29 14:53:02 2008 robUpdateComputersRebooting

Quote:
All of the computers are now showing green lights.

Remaining problems:

Alignment scripts are failing with "ERROR: LDS - NDS server error #13"
I think this is a server transmission error.

Dataviwer shows all channels as zero.


Fixed. Just started the testpoint manager on fb40m.


su
/usr/controls/tpman &
  593   Sun Jun 29 18:58:43 2008 ranaSummaryComputers1e20 is too big for AWG and/or IOVME
While testing out my matlab/awgstream based McWFS diagnostic script I accidentally put a
huge excitation into
C1:IOO-WFS1_PIT_EXC
. This went to 1e20 and then caused
some SUS to trip and c1susvme2 to go red. I tried booting it via the normal procedures
but it wouldn't come back, even after 2 crate power cycles. I also tried booting AWG
via the vmeBusReset, but that didn't do it. Then I booted c1iovme from the telnet prompt
and then I could restart c1susvme2 successfully.

The reason the excitation was so large is that the following filter command is unstable:
[b,a] = butter(4,[0.02 30]/1024);

The low pass part is OK, but it looks like making such a low frequency digital filter
is not. Que lastima. On the bright side, the code now has some excitation amplitude
checking.
  606   Mon Jun 30 16:00:02 2008 josephb, samConfigurationComputers 
Sam and I setup Cat6 cable from Megatron to the 1Y6 Switch (131.215.113.252) and also connected the 1Y6 Hub to the control room switch.

While I was at it, I checked the configurations of the two switchs now connected (one in 1X4 and one in 1Y6) to the martian network. For some reason, the 1X4 had switched to DHCP enabled and was using 131.215.113.105 as an IP address. I had thought I had setup it correctly initially, so am not sure what caused the change.

The easiest way I know of to check the setup is use smartwizard discovery program from the Netgear install CD (in the equipment manual file cabinet of the control room) on a windows machine. The passwords have been set to the controls password.

Megatron should now see and be accessible through the martian network.
  608   Tue Jul 1 09:26:33 2008 steveUpdateComputersRFM network is down
  610   Tue Jul 1 11:53:38 2008 YoichiUpdateComputersRFM network back
I took a tour of the FE machines and power cycled all of them.
After executing the software restart procedures of those computers, the RFM network got back to the normal state.
For some reason, the computers requiring startup.cmd (like c1lsc) halt after running this command. Actually the computer is running ok, but the command freezes. Basically, what it does is simply to load a kernel module. I don't know what is wrong.
Anyway, I just closed the terminal after running startup.cmd and it seems fine for now.
  614   Tue Jul 1 13:34:29 2008 robUpdateComputersRFM network back

Quote:

For some reason, the computers requiring startup.cmd (like c1lsc) halt after running this command. Actually the computer is running ok, but the command freezes. Basically, what it does is simply to load a kernel module. I don't know what is wrong.
Anyway, I just closed the terminal after running startup.cmd and it seems fine for now.


This is normal. On the linux RTFEs (Real-Time Front Ends), the real-time code totally hijacks the kernel, disallowing any interrupts. The system thus becomes totally unresponsive while the code is running, and communicates only through the RFM and the VME backplane.
  625   Wed Jul 2 17:19:03 2008 JohnSummaryComputersop440m - shutdown and restarted
After 160days op440m was getting a little slow.
  631   Thu Jul 3 13:54:26 2008 robConfigurationComputersmDV on rosalba

Does mDV work on rosalba? It can't find NDS_GetChannels. Looking on mafalda, I see that NDS_GetChannels is a mexglx. I think this means someone may need to compile it for 64-bit matlab before we can have mDV on rosalba. When that's done, we should get mDV running on megatron.
  636   Sun Jul 6 16:17:40 2008 tobinHowToComputersSVN
I was able to check out the 40m SVN here in Livingston using this command:

svn co svn+ssh://controls@nodus.ligo.caltech.edu/cvs/cds/caltech/svn/trunk/medm

As you might guess, this uses ssh in place of the web server (which we don't have yet).
  641   Mon Jul 7 14:02:05 2008 YoichiUpdateComputersSVN conversion progress
So far /cvs/cds/caltech/medm, /cvs/cds/caltech/chans and /cvs/cds/caltech/scripts have been converted to svn working copies.
Now /cvs/cds/caltech/target is being converted.
  644   Tue Jul 8 00:14:28 2008 JohnSummaryComputersAlarm handler
Rob thought it would be nice to have some alarms on the cpu loads and FE syncs.
I added all these channels to the alarm handler config file and wrote a script
which would set their values (HIHI,HIGH etc).

Ezcawrite allowed me to set the alarm levels (and ezcaread would give the correct
value) but no matter what I set the value to the alarm wouldn't sound.

After experimenting with a few other channels it appears that the alarm handler will
not show alarms if the alarm levels are absent from the db file (even though ezca
gives a value).

I edited the following files so we can have alarms on the cpus.

In c1iscepics:
lsc40m.db
asc40m1.db

In c1losepics:
bs.db
etmx.db
etmy.db
itmx.db
itmy.db
mc1.db
mc2.db
mc3.db
prm.db
srm.db

I saved backups in the appropriate folders.

Next time we have a bootfest please also do c1iscepics and c1losepics so these changes
will be implemented.
  654   Thu Jul 10 13:47:12 2008 YoichiHowToComputerssvn access via https
Now you can access to the svn repository on nodus by https.
To perform a checkout, you can use the following command

svn co --username svn40m https://nodus.ligo.caltech.edu:30889/svn/trunk/chans

This will check out "chans" directory.
The password for svn40m is written in the usual place.
You can also access the URL by a web browser to see the repository in a very primitive way.
A nice web interface for browsing the repository is planed but not yet implemented.
ELOG V3.1.3-