C&CZ Problem Card

Last night the Matlab license server crashed, because the number of open file handles reached the maximum configured, because a user started a lot of Matlab processes. Because we do not yet monitor this properly and it was not reported to us, we could only repair it this morning. We increased the number of file handles by a factor of four. We assume that this will be sufficient to wait till the reboot of next Monday, when it will again be enlarged, this time by a factor of sixteen. We will also on short notice start to monitor the Matlab license server properly. ...

The server lilo4 stalled at shutdown during the monday morning reboot. Solution: it continued its reboot after server wielewaal became available.

Quite a number of websites of ICIS were not available amongst which was eGroupWare. Furthermore, the boot process of workstations was impaired (because of the missing tftp). A reboot of the server seems to be sufficient to solve the issue. Affected websites: blackgem brainhealth croatia14 croatia15 cs eduspec eduspecorg egw ethergids fbkr kickstart mhealth oni snn

The DHCP server didn’t have free leases on Vlan136, so new devices on Vlan136 didn’t get an IP address, so didn’t have a working network connection. After we freed some unused numbers for PCs with a fixed address, we were able to drastically enlarge the number of leases on Vlan136.

In Huygens wing 2 a network switchmodule crashed twice. About 30 pc’s in the Study Area lost their connection to the network for some time. The switchmodule will be replaced on 2016-01-21 at 07:30 am.

Affected sites: aerochem akn aqua bb2syllabus bgard boekingen boekwur bomengids cac cafe celbi dossiers est2b exo filosofie fnwi hfml iconic ifl irmacrash itp2011 iwwrbook kanboard kndv labservant leonardo math mbfys mbp nieuwescheikunde nlt notebooks olympus pilab piwik printeger projecten puff pz qpl quantumcs refbase resourceanalysis rrl scheiolymp setup siam sigma-lustrum sito smartech4health studiereis sws tango taskboard tk toernooi tredd trooster ubuntumirror vsc w4kangoeroe wortel

The virtual hosts fret, wolf and vos received updates to libvirt packages, causing the network interfaces of the virtual machines running on them to fail. Solved after reboot (fret) or restart of the libvirt daemons and the virtual machines. Affected virtual machines: bfacdc03 bfacdc04 drukker est2 gitlab3 ms2 nwidc01 redmine2 smtp1 smtp2 ts2

The server lilo3/4 stalled at shutdown during the monday morning reboot. Solution: power cycle. Cause currently unknown.

The server comas1 did not start due to some possible timing problems between the raid array and OS. Solution: skipping the mount of /vd0 and /vd1 during boot (key S), mounts fixed manually later on followed by exportfs -uaf; exportfs -raf. Machine failed for unknown reason at boot. Solved by rebooting.