Frank's Activity Log

Admin

• Open issue 57834 with Unipress
  • Safari 2.0.2 does not work with a number of the menu items on the left frame
  • Unipress says this is a known issue -- they have a patch and will send it to me
  • Patch received and installed
• Put mac version of Firefox 1.5 on software download page and announced it
• meeting about future data center expansion

Posted by at 8:41 AM | Permalink | Comments (0)

Admin

• Request replacement for 146GB drive that is failing in zoosan5.
• Examine Oracle Calendar DB_VISTA Error 10 -- lock timed out.
  • Suggest that someone work with the client (9.0.4.1 windows machine) to get them upgraded
• rack2f filesystem full condition -- one ugrad moved
• configured blog webalizer stats

Posted by at 7:42 AM | Permalink | Comments (0)

Admin

• Realization that it is my fault (for not checking properly) that the ugrad's account I moved last Wednesday caused data loss.
• Directory permission problems in the Perkins project. Unable to find cause... Asked team for assistance
• Updated fake SSL certificates for test calendar server to stop the daily warning messages
• Updated moveuser.dce script to put a timestamp on the done log entry.

Footprints

• Migrated database to dedicated MySQL server to provide guaranteed availability while the corporate MySQL server is upgraded
• Note: I forgot to disable BB monitoring and got lucky that no pages were sent out!

Posted by at 8:10 AM | Permalink | Comments (0)

Footprints

• Upgraded the mysql libraries on the server to 4.1.5
• Discovered that the problem with the socket for clients to talk to the server was that the permissions on the directory were not allowing anyone other than the mysql userid to see the contents!
• DOH!!!

Posted by at 7:28 AM | Permalink | Comments (0)

Footprints

• Attempted upgrade of the server only partially successful
  • Local MySQL server -- FAILURE
  • 7.0d3 upgrade -- SUCCESS
  • RHEL3u6 upgrade -- SUCCESS
• The 6.5c footprints server was unable to communicate with the local MySQL server via the socket. Other things were able -- so, I'm certain that it's just a stupid error on my part.

Posted by at 7:19 AM | Permalink | Comments (0)

Admin

• Update permissions for Perkins project
• Sexual Harassment Training
• rack2d 97% full -- ugrad account uploaded 10GB of pictures

Posted by at 12:28 PM | Permalink | Comments (0)

Admin

• /tmp on cgi.uvm.edu -- overly full due to MySQL files left there over the weekend
• TELE replacement -- email conversation
• Footprints upgrade plan
• digital art departmental id having issues
  • Determined the home directory was owned by root:system instead of the correct ID
  • Corrected the problem
  • Unsure of why the problem existed
  • Sent email to team about the problem for input
• Added entry for giraffe samba for B. Hogel
• What happened to oldzoo.uvm.edu???
• perkins landscape files needed a different owner/group
• modify an ERP consultant LDAP entry so it gets into AD
• Create FootPrints upgrade checklist for Friday

Posted by at 9:11 AM | Permalink | Comments (0)

Admin

• Recycle sshd on penguin6 -- it's not working from moose again
• Create two accounts on penguins one from Friday and one from this morning that were not created due to the continuing problems with this sshd

Posted by at 5:32 PM | Permalink | Comments (0)

Admin

• Clean up account renaming failure from yesterday
• Recover a MS Access (.mdb) file that someone managed to corrupt
• name server problem with the in-mailgw record preventing some MX's from working
• TELE replacement meeting
• Sync_ML discussion
• RSA SecurID ... ignore root!

Posted by at 9:24 AM | Permalink | Comments (0)

Admin

• WebCT: assist application owner with installation of migration software
• Software: Assist Geoff with ownership of default.php file -- why does it have to be "waw" for things to work???
• MySQL: update sql_check script to include "mysqladmin status" output
• TSG meeting

Posted by at 9:23 AM | Permalink | Comments (0)

Admin

• LDAP: Further investigation and discussion with Registrar's office concerning a duplication of SSN numbers
• Propose a separate MySQL server for TSG production uses so the consistant problems caused by sandbox and web server accesses of the MySQL server stop making webmail and footprints unavailable.
• Copy the new ashelp.html file to the production footprints server (it was installed and tested on the test server yesterday -- apparently I forgot to blog anything yesterday)
• Notified the remaining folks that they need to update and test their group.html file equivalents before FootPrints gets upgraded on Nov 25
• FP ISSUE=17129: Hung IMAPD -- removed
• Code review for Geoff
• UPS Meeting in MANN
• Fix path bug in account activation web page

Posted by at 9:42 AM | Permalink | Comments (0)

LDAP/Accounts

• BANNER feed contains four entries that get SSN values added and with that additional information the LDAP Update process determines there are duplicate records in the Directory.
  • two of them are previous Student Employee errors that I was unable to find a match for in the Student informations
  • One is a collision with an expired entry from July 2004 that was to have gone away next summer
  • One is a collision with an existing entry that already has a BANNER and HRS component -- got the registrar's office looking into that one

Posted by at 11:21 AM | Permalink | Comments (0)

Admin

• FP Issue=16731 Proj=4: gave directions for using the Address Book in Thunderbird to look people up in the UVM LDAP Directory.
• FP Issue=17022 PROJ=4: recovered webmail addressbook (MySQL table records) from October 30 backup
• FP ISSUE=17066 PROJ=4: hung imapd on penguin2 -- removed
• FP ISSUE=17070 PROJ=4: hung imapd on penguin5 -- removed
• Consultant account: updated LDAP directory so it will get properly copied into AD
• FP ISSUE=17075 PROJ=4: hung imapd on penguin5 -- removed
• FP ISSUE=17082 PROJ=4: hung imapd on penguin4 -- removed
• Create requested blog for a user

Footprints

• Shutdown the production server (5am - 6:15am) and copied the database/code to mink for testing purposes
• Actual copy time was less than five minutes!
• Had to upgrade DBD-mysql to the version from 7.0 to get it to work with the MySQL 4.1 database
• Upgraded to 7.0d3 -- caused a failure.
  • Had to manually create the db/ABMASTER1/MR/abmasterMasterTableCreated file to get around the 203 error when administrator account logged in
• Announce the availability of the test server...

MySQL

• rewrite the sql_check.bash script for the test server

Posted by at 7:10 AM | Permalink | Comments (0)

Admin

• FP Issue=17016 proj=4: Hung imapd -- responded to SigUSR1 -- Killed
• FP Issue=17021 proj=4: Lost email -- recovered INBOX from day before
• FP ISSUE=17022 PROJ=4: Hung imapd -- responded to SigUSR1 -- Killed
• Remove old SSH.com 2.4.0 code from otter
• FP ISSUE=17022 PROj=4: lost webmail address book
  • Working with Scott to identify which table to look at
  • turba_* is the list of tables to look at
  • Recovered backups from 11/7 and 11/1 and did not find any addressbook entries for the specified NetID in either copy
  • Updated issue with need to know when the addressbook was last known to exist so we can recover it
• rack2h space issue... walk-in uploaded 1.5GB to backup a machine during a rebuild... properly chastised!

FootPrints

• Installing 7.0 on test server
• Oops... DBD is looking for mysql_config and we don't have it!
• Ok... got it... had to install mysql-devel package and gcc...
• Scheduled downtime for production system for 5am to 7am Friday morning

Posted by at 7:23 AM | Permalink | Comments (0)

Admin

• On Call day
• SMmonitor stopped running on ermine. Paged at 0700. It stopped somewhere between 0630 and 0700. Restarted. It stayed running -- assume it is working.
• LDAP
  • Registrar fixed the duplicate PIDM problem from Saturday by deleting the one that LDAP had created an account for (Ugh!).
  • Reclaimed the expired entry with the remaining PIDM information.
• FP Issue=17001 Proj=4: Hung IMAPD, responded to SigUSR1, killed!
• rack1b 97% full
  • moved the big user to clear space
  • Culprit was edml177 which uploaded 1GB of class materials
• DNS update for ResLife machine
• MySQL hung up, all clients busy doing something, kill -9 required to get it to restart

SecurID

• Catalyst FP Issue=599 -- SSH.COM not working
  • Fixing issue #1 -- my inability to log into cardinal at all
    • Seems my account didn't have a home directory
    • didn't have a home directory
    • And didn't have a valid shell
    • Amazingly, the computer wouldn't let me log in (GRIN)
  • Brief sidetrip to investigate PAM support in sudo -- determined that since sudo has not been tested with AIX PAM, that this is a good road to not go down at this time.
  • Giraffe fixed by adding "keyboard-interactive" to the list of AllowedAuthentications.
  • Otter's version of the SSH.com client is too old to support the "keyboard-interactive" Authentication method. Should we remove it?
  • Windows SSH.COM ssh client needs to be at least 3.2.x and needs to have keyboard interactive set as a valid authentication mechanism.
  • Have to be careful, because upgrading from an older version doesn't appear to replace the settings file and therefore the newer default authentication methods don't get added
  • Issue closed

Calendar

• TAR 4787052.993
  • Soft closed -- waiting for recurrance
  • Updated the csmon.sh script to keep seven days worth of info to avoid filling up the disk

Posted by at 7:35 AM | Permalink | Comments (0)

Admin

• Walk ces rep through problem with PHP web page
• Created a blog for a user who requested it on Saturday
• LDAP
  • Duplicate PIDM entry in Saturday's update
  • It truly is a duplication... Registrar's office notified
• FP issue=16985 proj=4: Hung imap... sigusr1 responsive... killed

Calendar

• TAR 4848713.992 -- PPC browser
  • Official answer is Oracle does not support the PPC Browser that SyncML is the way to go with these devices
  • The offered to enter an enhancement request
  • I responded that BUG 3662237 seems to be the same issue and that guilty until proven innocent (broken until certified to work) is not a very defensible position
  • Oracle is opening bug 4722248 as an Enhancement Request (ER) for this topic.
• TAR 4787052.993 -- db_Vista 10 during backup
  • Oracle updated with source for fedora core 3 procinfo tool
  • Will download and attempt to compile on RHEL3u5
    • Requires libtermcap-devel for building
    • Seems to build cleanly after that
    • Installs and runs on test and production systems
    • monitoring script set to run every 20 minutes on production system

FootPrints

• 56359:
  • UniPress has not updated this FP since 2. November and I can find no evidence they have accessed the system to check anything out.
  • Calling to speak with a rep to see what is going on.
  • Concerned with support for 6.5c being dropped tomorrow.
  • Made their account project admin for the project with the problem
  • Thirty minutes on the phone and Dan was able to find and fix the issue
  • MRusers files for idividuals did not list an agent type (2) as the second colon delimited field
  • projectLogins files did not list the projects which had the agent type missing from the MRusers files of the same name

Posted by at 8:08 AM | Permalink | Comments (0)

Calendar

• TAR 4776115.993
  • Information from oracle support confirms that this problem was caused by the specific mailignoreerror that was the recommended fix for TAR 4560587.992 is the direct cause of this incident.
• TAR 4848713.992
  • Oracle says that the Pocket PC browser is not supported and soft closed the TAR!
  • I responded that their response was not satisfactory

Admin

• Giraffe /var overly full. Deleted syslog file (135MB) from 29 November 2004.
• FP Issue-16949 proj=4 -- hung imapd
• Mink: Install fresh copy of RHEL3u6 for FP7 and MySQL4.1 testing
• UPS Trap forwarding -- New list from the windows weenies of machines to forward the traps to
• SecurID:
  • Cardinal becomes a client
  • ERP Firewall update info passed to ERP folks again
• Reboot lion ... 744 days uptime ... way slowed down...
• Reboot didn't fix lion... must be the UPS itself that is so slow

Posted by at 8:08 AM | Permalink | Comments (0)

FootPrints

• 56359
  • reset the unipress footprints account password to a shared secret so they can log in to check things out

Calendar

• TAR 4776115.993
  • Oracle wants me to add the following line to the [CWS] section:
    mailignoreerror = {64,68,71,75,67,70}
    This will add errors 64 and 68 to the default list of 71, 75, 67, and 70. The present entry has just {64} -- which was the recommendation from TAR 4560587.992... was this problem caused by Oracle???
• Unsupported browser
  • Research issue with browser from Axim X5 and Audiovox Thera getting sent to a "Your browser is not supported" page
  • This appears to be an instance of Oracle bug 3662237
  • Oracle has implemented a list of known to work browsers and all other browsers are assumed not to work
  • Open TAR 4848713.992 for this

Admin

• Watch machine room for A/C temporary unit testing
  • Orientation of the temporary units discharge port makes a remarkable difference in our automated sensor readings.
  • The one on the east wall being angled toward the sensor causes it to read a good humidity level and a cool temperature
  • Definitely airflow patterns...
• test horde-3.0.6 version of webmail
• Remove mink monitoring from BB
• Team meeting

Posted by at 7:30 AM | Permalink | Comments (0)

Admin

• FP Issue=16903 proj=4: Hung imapd
• FP Issue=16904 proj=4: Hung imapd
• FP Issue=16906 proj=4: Hung imapd
• groundhog smtp measurements...
  • Looks like groundhog was shutdown yesterday -- will verify that later with Jim.
  • discovered that whistlepig is running a newer version of sendmail than the scripts were written to handle so updated the scripts to deal with both the new and old versions of sendmail statistics
• Applied 10.4.3 update to laptop
• Downloaded and burned IBM's UpdateXpress CD 4.02 version -- 2005/10/10 release
• run Oracle root install script for MJG on cat000

WebCT 6.0

• Installed RHEL3u6 on carp so I can keep it updated and ready for when webct gets installed

FootPrints

• Issue 56359: Unipress provided a new proj30 file which was installed but doesn't seem to fix the problem. Waiting to hear whether it requires the footprints web processes to be cycled or not.

Posted by at 10:31 AM | Permalink | Comments (0)