[SOLVED][Storage & Snapshots] Failed to expand volume "$Name". Storage pool: 1. Unable to resize file system.

[Marcrogers]

Using a TS-831 on various firmwares - currently on QTS 5.0 - I don't think its a firmware issue.

After performing a "one by one" disk upgrade and then using the built in volume resize option "Manage Storage & Snapshpts" I kept running into this error when resizing my volume: "Storage & Snapshots Volume [Storage & Snapshots] Failed to expand volume "DataVol1". Storage pool: 1. Unable to resize file system."

Adding to my frustration it also kept leaving the filesystem dirty, and attempts to run e2fsck from the GUI proved futile.

The following posts were helpful in putting 2 & 2 together:
viewtopic.php?t=123461
viewtopic.php?f=25&t=9083
viewtopic.php?t=12109

First I had to solve the e2fsck problem. It turns out that in most cases this was being caused by open file handles attached to the raid volume despite it being unmounted. so presumably the GUI attempts to unmount, runs the check and then fails because of those file handles. Using LSOF and killing anything still attached to the volume allowed me to run "e2fsck /dev/mapper/cachedev1" without any issues

Second thing I noticed is that using
/etc/init.d/services.sh stop
/etc/init.d/opentftp.sh stop
/etc/init.d/Qthttpd.sh stop
/etc/init.d/rsnap.sh stop

to stop all services does not do it cleanly, there are numerous errors outputted to stdout and its clear that some child processes are not cleaned up leading to the file handle issue. QuLog appears to be one of the main culprits.

one of the posts I linked above suggested halting QPKG and disabling all services that you can. Just disabling all the apps one by hone helped hugely. Then check for any orphaned processes and check with lost for anything linked to your MD device or /dev/mapper/cachedev1.

at this point you can cleanly unmount /dev/mapper/cachedev1 and run another e2fsck. once thats done I was able to go back into the GUI and tell it to resize the volume. No need to remount anything, it just worked.

my guess is that the e2fsck and volume resize functions regularly fail because they are unable to cleanly unmount the volume.

anyway, hope this helps someone so that my weekend wasn't entirely wasted.

Marc