Opened 7 years ago

Closed 7 years ago

#5744 closed Bug/Something is broken (wontfix)

kernel upgrade of negri and malaka

Reported by: Owned by:
Priority: Medium Component: Tech
Keywords: server-upgrades Cc:
Sensitive: no


This ticket documents the upgrade of negri and malaka:

The first thing I noticed with the upgrade process is that negri has a servers listed at:


however, I don't have access to this server and it's not in puppet. I would like to know what the situation is with this server.

Change History (12)

comment:1 Changed 7 years ago by

We still seem to have the dely in response after taking down a server. It took about 15 seconds for negri to respond after issuing sv down to all the guests.

comment:2 Changed 7 years ago by

clyde is requesting a password with the following output:

  Volume group "clyde" not found
  Skipping volume group clyde
Unable to find LVM volume clyde/root
  Volume group "clyde" not found
  Skipping volume group clyde
Unable to find LVM volume clyde/swap_1
Unlocking the disk /dev/disk/by-uuid/95e54a59-a2d1-4b90-995f-218041b7add5 (hda2_crypt)
Enter passphrase:

comment:3 Changed 7 years ago by

We still have a blank screen for reboots on negri's guests. That is there is no grub menu display for the guests.

comment:4 Changed 7 years ago by

mx3 failed fsck saying that /tmp and /var were already mounted. log file is saved in /var/log/fsck/checkfs

comment:5 Changed 7 years ago by

mx1 is showing a bunch of console lines like this one:

[  653.192747] [FIAIF_MARTIAN]:IN=eth0 OUT= MAC=02:09:00:00:00:02:00:19:2f:e8:fa:00:08:00 SRC= DST= LEN=60 TOS=0x00 PREC=0x00 TTL=45 ID=2249 DF PROTO=TCP SPT=33940 DPT=80 WINDOW=5840 RES=0x00 SYN URGP=0 

I don't know what this means or if we need to worry about it.

comment:6 Changed 7 years ago by

  • Owner set to
  • Status changed from new to assigned

Jamie please review.

comment:7 Changed 7 years ago by

  • Owner changed from to

Unfortunately, it looks like I made a relatively large error in the upgrade process. Using cluster ssh to perform the necessary upgrades, I seemed to have failed to actually upgrade the servers. Looks like I will get to do this all over again.

comment:8 Changed 7 years ago by

clyde is managed by aktivix - and they can add their own passphrase, so that is perfectly fine.

The mx1 lines are a result of a special firewall installed on mx1. Also ok to ignore.

The errors on mx3 are due to my screwup when I sync'ed from the mx3 in canada to our server (I didn't use --delete), which resulted in duplicate files in /etc/rc2/ - which causes file systems to be mounted twice.

I'm working on cleaning that up now.


comment:9 Changed 7 years ago by

There are still some duplicates in /etc/rc*/ but I think I cleaned out the duplicate mount ones from /etc/rcS.d so the next boot should be clean.

comment:10 Changed 7 years ago by

I also ran into these errors on mx1. On trying to configure dependency based boot, mx1 failed. This was configuring sysv-rc. Most of the errors said something like "missing LSB tags and overrides." For some reason I couldn't cut and paste the output of the cssh bash window. So it was difficult to include all of the errors.

However, I believe they can be repeated by running dpkg-reconfigure sysv-rc.

comment:11 Changed 7 years ago by

  • Owner changed from to

All guests on negri and malaka are now up to date as are the hosts.

Jamie please check comment:10 regarding mx1.

comment:12 Changed 7 years ago by

  • Resolution set to wontfix
  • Status changed from assigned to closed

I'm going to suggest won't fix on the mx1 errors. Once we finish the transfer of mx25 to malaka, we will start on the process of moving the mexican members into the MFPL control panel, which will involve moving them off the mx? servers, meaning that mx1 will be retired. Even thought it's a ways away, I think it's acceptable to put off a switch to dependency based boot.

Please login to add comments to this ticket.

Note: See TracTickets for help on using tickets.