Did a quick breakdown of the costs of a simple Website development project for LP, who is submitting a SSHRC application.
August 19, 2008
HP came in with her class, and among other things needed to do the telephoning activity. When we started up the Sanako teacher application, we got the same RSRV error as the other day. I was still hitting the same problem after starting RSRV.exe manually several times. Eventually, after confirming that both RSRV.exe and RSRVAny.exe were running, I turned off the console and turned it back on, then restarted the application; that finally did it. My guess is that RSRV.exe has to be running from boot time, not logon, otherwise it has trouble connecting to the console.
Even after everything was running, and we switched Group A to the Telephone activity, we had trouble getting stations to connect to each other; stations which were not connected to anyone were still reporting themselves as Busy when called by another station.
This teacher station is pretty much unusable without one of us physically in the room to get it going, and even then it takes a while and is flaky. I wonder if it might be better to roll it back from Server 2003 to Server 2000, like Lab C, which as far as I know is working OK.
Late yesterday afternoon, HP called me from the ELC about her class today, in Lab B. She wants to show a video. I checked the TV cart in Lab B this morning, and it has no VCR on it (has it been stolen, or was it never there?). I decided I'd move the one from lab C, which is set up and working, but the cart has been padlocked to one of the desks, so that's out. There is another TV+VCR cart in the materials room, so I've rolled that into Lab B in the hope that it will work.
On starting up the servers this morning, I noted the following problems:
- The mouse doesn't work on the KVM (on any channel).
- Squash is full of dust.
- When Squash goes through a cold boot, it gets stuck repeating the memory check, so you have to be there to hit the ESC key to make it abandon the check and continue to boot.
- Both Squash and Parsnip had outstanding Windows updates, so I did those.
- Quince had login problems, which I've discussed with Greg.
I'm going to have a go at fixing the mouse problem -- it may just be a connector, or it could be that one of the KVMs is dying. The Quince problems we think we've solved, and the rest will wait till Greg gets back.
August 18, 2008
I'm the only one in this week too, it turns out, so I'll be racking up the G&T hours this week.
Unplugged everything in Lab C again for tonight's outage, then started on the servers. Shut down Chou by SSH. Couldn't log in to Quince at all, either as root or hcmc, with the given pwds; I tried this both remotely and on the KVM in the server room. Finally I had to shut down the server with the power switch. Shut down Parsnip with no problem on the KVM. When I came to Squash, I discovered that it had actually failed to start properly this morning, and appeared to be locked in a loop going through its memory test. Restarted it a couple of times, eventually skipping the memory test, but when it came up it reported one or more services had failed. Shut it down. This will need looking at tomorrow.
Still waiting to shut down Idaho because I'm doing ColDesp backups at the moment.
JT came by with substantial corrections to vol 17 articles, which I'll try to get done this week. At the same time, we created, uploaded, and linked a PDF of the Notes for Contributors, which was not previously available on the site.
Today I built the image browser page. I also ran another linking XQuery job on the database after realizing that many of the VI scans were actually from 1857 and 1859, not just from 1858, so a couple of hundred more documents are now linked to images. After that, I went through every single document in the range of the digitized microfilms, and checked that it was in fact linked correctly where an image exists. A few had been missed, usually because there was no page reference in the metadata. I also discovered that the CO 60 3 scans had been named inconsistently with the others -- the page numbers were done at four digits instead of five -- so I renamed all of those files in all locations on the server and on my drive, and relinked those documents.
Once everything was working, I rewrote the Development page on the site to reflect the current state of play, created an introductory page for the page image browser, and then re-jigged the site menu to add the page image link. To make space, I removed the link to the blog from the main menu; it's linked from the Development page anyway, and it's not of interest to most visitors.
I think I've now done as much as I can with the page images, until there's manpower to link each individual image to a <pb> tag in the transcription. Next will be linking names to bios.
Powered everything back up after the weekend outage with no problems; testing sending out a file in Lab C. All seems to be OK.
This post does not include changes to my vacation; it is for accounting purposes only.
As of this coming Friday I will have taken 15 of my 23 days, leaving me with 8 days vacation to take.
Days I took:
July 14-25 (10)
August 11-22 (10)