Auto-extend not working correctly

Question

Hi,

Has anyone else had problems with the auto-extend feature on Celerra? We set this up with virtual provisioning to try and save some disk space. Unfortunately on one filesystem the filesystem didn't extend correctly and so the it filled up preventing the users from saving files.

Now our users don't trust the system and are insisting that we allocate them actual space rather than using vp.

EMC couldn't tell me why the auto-extend didn't work correctly and it has functioned ok since the problem. However users can't take the risk that this will happen again and prevent them from working. It is annoying as I lose the benefit of virtual provisioning.

Sheppie

richcrompton · Answer

I am using thin provisioning on an NS40 (integrated, v5.5) with auto-extend enabled. Can you post the output of 'nas_fs -info ' (where is the name of the virtually-provisioned file system)? Also include entries from the output of 'server_log server_2' around the time the file system reached capacity.

sheppie · Answer

I think this is the relevant part of the output from nas_fs:

id = 637
name = fs11a
acl = 0
in_use = True
type = uxfs
worm = off
volume = v1002
pool = clarata_archive
member_of = root_avm_fs_group_10
rw_servers= server_2
ro_servers=
rw_vdms = vdma1
ro_vdms =
auto_ext = hwm=90%,max_size=512000M,virtual_provision=yes

Unfortunately the logs have rolled over due to a datamover failover caused by a bug in the OS. This is one of the reasons why EMC couldn't explain why the auto-extend didn't work.

The reason I asked if anyone else has seen this problem is that I am worried that it will happen again.

Sheppie

richcrompton · Answer

Our NS40 is running v5.5.31.6 and I haven't had any auto-extend issues since it was installed 5 months ago. Two differences between your FS and mine, though: 1) I'm not using a VDM and 2) all of my NAS file systems are on RAID-5 FC RAID groups.

What about running the following for the relevant file system:
cat /nas/bin/cmd_log* | grep "rootnas_fs -x" | grep

or:
cat /nas/log/nas_log.al.err | grep

and one more:
cat /nas/log/sys_log* | grep "fs auto extend" | grep

Each of these sets of logs on my NAS has information related to file systems auto extensions, the "nas_log.al.err" with the most helpful root cause data, I believe.

dynamox · Answer

i had the same issue on my NS80 ..file system was virtually provisioned and set to auto-extend ..i started dumping data into it and all of sudden my DM panicked. I never got reply from support why auto-extend did not work but i did get reply why DM panicked. It was something about file system size that was created and they had to use utility called ckfs to fix/identify problem file systems, DART bug i assume.

Celerra

Auto-extend not working correctly

Was this post helpful?