Opened 10 years ago

Closed 8 years ago

#457 closed defect (fixed)

pxenodes hang on boot, pdnsd not started on head node

Reported by: Fitz, Skylar Owned by: skylar
Priority: major Milestone:
Component: Liberated Version:
Keywords: Cc:
Blocked By: Blocking:
Estimated Hours: 0 Total Hours: 7.25

Description (last modified by mmludin08)

Sometimes PXE-booted nodes will hang just after the message "Opening socket /var/cache/pdnsd/pdnsd.status" (possibly unrelated).

They often need 2+ reboots until the login prompt is reached.

This is most likely due to a race condition in the locking between NFS and AUFS.

Merged duplicate ticket #611 below -- /* Mobeen */ Please refer to #611 for work done so far

pdnsd not started on head node

need these changes:

/etc/default/pdnsd:

START_DAEMON=yes

/etc/pdnsd.conf:

server_ip="192.168.3.1"

Change History (38)

comment:1 Changed 10 years ago by skylar

unfortunately updating to the latest aufs with Linux 2.6.31.12 doesn't seem to solve this problem

comment:2 Changed 10 years ago by fitz

  • Component set to Liberated
  • Description modified (diff)

comment:3 Changed 10 years ago by skylar

  • Owner changed from somebody to skylar
  • Status changed from new to assigned

this might be getting worse - having trouble with the new 2.6.31.12_aufs kernel

comment:4 Changed 10 years ago by skylar

got git figured out, checking out 2.6.34 onto t-voc

comment:5 Changed 9 years ago by skylar

0.5 hours logged for skylar: git commands

check out tree:

git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git

list tags:

git tag -l

switch tags:

git checkout -b v2.6.34 v2.6.34

comment:6 Changed 9 years ago by skylar

0.5 hours logged for skylar: hardware finally cooperated to do builds

comment:7 Changed 9 years ago by skylar

0.2 hours logged for skylar: restarting 32-bit and 64-bit builds

comment:8 Changed 9 years ago by skylar

trying another update

comment:9 Changed 9 years ago by skylar

0.5 hours logged for skylar: building kernel with debug code

comment:10 Changed 9 years ago by skylar

1.3 hours logged for skylar: getting errors with 2.6.35, rolling back to 2.6.31.12 and rebuilding with debugging on

comment:11 Changed 9 years ago by skylar

0.5 hours logged for skylar: rolling back merges

comment:12 Changed 9 years ago by skylar

pulling down new kernel

comment:13 Changed 9 years ago by skylar

0.3 hours logged for skylar: trac pull still going

comment:14 Changed 9 years ago by mmludin08

  • Description modified (diff)
  • Reporter changed from fitz to Fitz, Skylar
  • Summary changed from pxenodes hang on boot to pxenodes hang on boot, pdnsd not started on head node

comment:15 Changed 9 years ago by mmludin08

  • Description modified (diff)

comment:16 Changed 9 years ago by skylar

In [3175]:

THIS IS A HACK - installing new kernel from /root (#457,#629)

comment:17 Changed 9 years ago by skylar

In [3176]:

dpkg --root breaks if architecture is different from host (#457,​629)

comment:18 Changed 9 years ago by skylar

In [3177]:

typo (#457,#629)

comment:18 Changed 9 years ago by skylar

In [3178]:

changing hard coded kernel version (#457,#629)

comment:19 Changed 9 years ago by skylar

In [3179]:

test name for fcopy (#457,#629)

comment:20 Changed 9 years ago by skylar

0.8 hours logged for skylar: more work getting kernel working

comment:21 Changed 9 years ago by skylar

In [3197]:

adding 3rd party kernel modules (#457,#629)

comment:22 Changed 9 years ago by skylar

In [3206]:

re-doing modules (#457,#629)

comment:23 Changed 9 years ago by skylar

0.5 hours logged for skylar: finally got bccd-ng3 working by virutal power cycling, trying new build w/ fresh modules

comment:24 Changed 9 years ago by skylar

0.5 hours logged for skylar: still playing around with amd64

comment:25 Changed 9 years ago by skylar

In [3218]:

new cloop (#457,​629)

comment:26 Changed 9 years ago by skylar

In [3219]:

updating kernel version number (#457,#629)

comment:27 Changed 9 years ago by skylar

In [3220]:

removing old grub (#457,#629)

comment:28 Changed 9 years ago by skylar

In [3221]:

just have default menu, still need bccd branding (#457,#629)

comment:29 Changed 9 years ago by skylar

In [3222]:

2.6.31.12 lingering (#457,#629)

comment:30 Changed 9 years ago by skylar

In [3223]:

adding Id (#457,#629)

comment:31 Changed 9 years ago by skylar

In [3224]:

changing symlinks during liberation so update-grub works (#457,#629)

comment:32 Changed 9 years ago by skylar

In [3230]:

merging in newer kernel, squashfs (#457, #629)

comment:33 Changed 9 years ago by skylar

In [3231]:

accidentally removed $release (#457,#629)

comment:34 Changed 8 years ago by skylar

In [3232]:

removing old manual install code (#457,#629)

comment:35 Changed 8 years ago by skylar

0.5 hours logged for skylar: issue seems to have gone away

comment:36 Changed 8 years ago by skylar

  • Status changed from assigned to qa

comment:37 Changed 8 years ago by skylar

  • Resolution set to fixed
  • Status changed from qa to closed

all good

Note: See TracTickets for help on using tickets.