Tag: linux

MongoDB Fails Updating on Debian

Every so often there’s a MongoDB update on my Debian VPS that fails. The output of ‘aptitude full-upgrade’ is:

# aptitude full-upgrade
The following partially installed packages will be configured:
mongodb-10gen
No packages will be installed, upgraded, or removed.
0 packages upgraded, 0 newly installed, 0 to remove and 0 not upgraded.
Need to get 0 B of archives. After unpacking 0 B will be used.
Setting up mongodb-10gen (2.0.5) …
Starting database: mongodb failed!
invoke-rc.d: initscript mongodb, action “start” failed.
dpkg: error processing mongodb-10gen (–configure):
subprocess installed post-installation script returned error exit status 1
configured to not write apport reports
Errors were encountered while processing:
mongodb-10gen
E: Sub-process /usr/bin/dpkg returned an error code (1)
A package failed to install. Trying to recover:
Setting up mongodb-10gen (2.0.5) …
Starting database: mongodb failed!
invoke-rc.d: initscript mongodb, action “start” failed.
dpkg: error processing mongodb-10gen (–configure):
subprocess installed post-installation script returned error exit status 1
Errors were encountered while processing:
mongodb-10gen

The update works fine, but mongo just fails to start properly.

The problem in my case is simply that there’s a /var/lib/mongodb/mongod.lock file lying around from some previous process. Deleting that file and re-running the aptitude command will start it properly. (Reminder post because I keep forgetting what the problem is.)

Flushing the Linux System Cache

Found this interesting page on the Linux memory management site which details how to clear the system cache in Linux. This was handy for us as we were testing file download performance off various VPS configurations and we wanted to be sure the file we were downloading was getting read off the disk, so we could more accurately measure the usual performance.

Basically you can just do: echo 3 > /proc/sys/vm/drop_caches and it will ‘free pagecache, dentries and inodes’. This worked a treat; we were able to download a file, then download it again (observing that the second download was much, much faster because it was coming out of the system cache), then run that echo and download again and observe that the speeds were back to the first download.