Break reports

Here are listed all breaks in HIIT's IT services.

Breaks in several servers at 2009-09-16 08:00-10:00

Description: 

Schedule:

2009-09-16 08:00 - 10:00

Duration:

2:00 h

Affected services:

LDAP, Radius (incl. eduroam), VPN.

Reason:

Kernel upgrade.

The following servers will be rebooted: find, lose, openvpn01.

Break in each server is approximately 5 minutes.

Update 08:17: Break is over.

Break in storage area network (SAN) at 2009-09-07 13:55-18:30

Description: 

Schedule:

2009-09-07 13:55 - 18:30

Duration:

4:35

Affected services:

At least VCS, Wiki, WWW, Windows AD. Possibly others.

Reason:

Problems with SAN.

Update: Problems were caused by faulty SFP (fibre adapter).

It caused one blade-enclosure's one uplink-port to fluctuate up and down thus messing up the SAN-fabric using that port. And, of course, availability of paths to LUNs and other resources accessed via that fabric fluctuated too. Because the port wasn't down all the time, determining the cause of problem wasn't clear, paths sometimes worked and sometimes didn't. Because of this, an incorrect desicion of rebooting one of the SAN-swithches, the one in the functioning fabric, at 14:18:35 caused some of the hosts to temporarely lose all paths to LUNs in SAN.

Linux is especially cranky when it comes to losing it's disks, even temporarely. It determines the disk to be read-only quite soon, even the disk would come back in a few seconds. To gain read-write access, a reboot was required.

The faulty SFP started to be down longer periods of time later and thus was discovered and replaced with a working one. Breaks to services were between 9 and 44 minutes. Kernel updates were installed to servers that needed them during the breaks as well.

Break in shell at 2009-08-27 08:00-08:30

Description: 

Schedule:

2009-08-27 08:00 - 08:30

Duration:

0:30 h

Affected services:

HIIT's general purpose server.

Reason:

Shell will be rebooted due to a kernel upgrade.

Update at 8:19: Break is over.

Break in version control (HIIT-VCS) at 2009-07-17 12:00 - ∞

Description: 

Schedule:

2009-07-17 12:00 - ∞

Duration:

Indefinitely

Affected services:

HIIT-VCS

Reason:

HIIT VCS will no longer be available by name vc.hiit.fi. Only working name will be vcs.hiit.fi

Name change was announced 2009-03-05. For more information please visit HIIT wiki at page Version Control System name changed

UPDATE: vc.hiit.fi was dropped 2009-07-21 10:25

Break in DHCP service at 2009-06-18 18:00-19:00

Description: 

Schedule:

2009-06-18 18:00 - 19:00

Duration:

1:00 h

Affected services:

DHCP service

Reason:

Operating system upgrade from Debian 4.0 to 5.0 and firmware updates.

Several individiual breaks during one hour. Each break will be shorter, max. 10-15 minutes each.

Update 18:56: Break is over.

Pages