Unsolved
This post is more than 5 years old
112 Posts
0
1245
VNX for file - nfs - snapshots
Hi, VNX for file, nfs share mounted on a vdm.
Automatic snapshots scheduled at .00 every hour.
From VNX I can see snapshots as expected:
but from a linux client mounting that file system I can see some snapshot which are not on VNX (!!) - highlighted in bold.
cheesep@XXX:~/.ckpt$ ls -alt
total 257
dr-xr-xr-x 2 root root 512 May 26 17:16 .
drwx-----x 3 cheesep staff 1024 May 22 14:17 2015_05_26_09.00.03_Europe_London
drwx-----x 3 cheesep staff 1024 May 22 14:17 2015_05_26_10.00.00_Europe_London
drwx-----x 3 cheesep staff 1024 May 22 14:17 2015_05_26_11.00.02_Europe_London
drwx-----x 3 cheesep staff 1024 May 22 14:17 2015_05_26_12.00.01_Europe_London
drwx-----x 3 cheesep staff 1024 May 22 14:17 2015_05_26_13.00.03_Europe_London
drwx-----x 3 cheesep staff 1024 May 22 14:17 2015_05_26_14.00.04_Europe_London
drwx-----x 3 cheesep staff 1024 May 22 14:17 2015_05_26_15.00.01_Europe_London
drwx-----x 3 cheesep staff 1024 May 22 14:17 2015_05_26_16.00.03_Europe_London
drwx-----x 3 cheesep staff 1024 May 22 14:17 2015_05_26_17.00.03_Europe_London
drwx-----x 3 cheesep staff 1024 May 22 14:17 2015_05_26_17.11.23_Europe_London
drwx-----x 3 cheesep staff 1024 May 22 14:17 2015_05_26_17.15.23_Europe_London
These last two "strange" snapshots can vary from a specific time to the next run of a snapshot job.
Any idea what can I check on the client side? Linux Ubuntu and Solaris clients behave the same.
EMC support already involved and nothing found on the VNX side.
Thanks.
matteo_marchett
112 Posts
0
May 26th, 2015 09:00
Yes it is. It replicates with another VNX at a secondary site every 1 minute.
If they are snapshots created by the replicaiton job, I'd expect their timestamp to change every 1 minutes or so, but it does not.
dynamox
1 Rookie
1 Rookie
•
20.4K Posts
0
May 26th, 2015 09:00
is that file system replicated ?
dynamox
1 Rookie
1 Rookie
•
20.4K Posts
0
May 26th, 2015 10:00
same replication schedule configured for vdm ?
Peter_EMC
674 Posts
0
May 26th, 2015 23:00
The timestamps would only change every minute, if there is data to be copied to the destination and if the amount of data to be copied to the destination could be finished in one minute.
You can check the checkpoints doing a
fs_ckpt fsname -l -a on the controlstation. Typically the first 2 with the name beginning with "root_rep_ckpt" belong to the replication session.
Verify this output with the output of the linux client (ls -alt)
cadencep45
1 Rookie
1 Rookie
•
299 Posts
0
May 27th, 2015 00:00
if they are replication snapshots, from the control station, running
nas_replicate -list
should list the replication sessions and then running
nas_replicate -info , where session name is provided from the output of the first command tells you if the replication session is up to date and the rate of data transfer between primary and secondary sites.
matteo_marchett
112 Posts
0
May 27th, 2015 02:00
Hi Peter, I don't fully understand the "connection" between these two:
[nasadmin@C-VNX1-CS ~]$ fs_ckpt c-nfs-staff -l -a
id ckpt_name creation_time inuse fullmark total_savvol_used ckpt_usage_on_savvol
754 root_rep_ckpt_301_700035 01/07/2015-16:52:00-GMT y 90% 20% 1%
755 root_rep_ckpt_301_700035 01/07/2015-16:52:03-GMT y 90% 20% 1%
and
cheesep@XXX:~/.ckpt$ ls -alt
drwx-----x 3 cheesep staff 1024 May 22 14:17 2015_05_27_09.00.04_Europe_London
drwx-----x 3 cheesep staff 1024 May 22 14:17 2015_05_27_10.00.00_Europe_London
drwx-----x 3 cheesep staff 1024 May 22 14:17 2015_05_27_10.00.33_Europe_London
drwx-----x 3 cheesep staff 1024 May 22 14:17 2015_05_27_10.11.53_Europe_London
Peter_EMC
674 Posts
0
May 27th, 2015 03:00
the fs_ckpt fsname -l -a is listing you all existing checkpoints of a FS.
The ip-replication is based on two checkpoints, they are refreshed round robin.
fs_ckpt is listing the creation time of the checkpoints, while the ls -l is giving the name based on the last uptime.
With a running ip-replication. fs_ckpt fsname -l -a will list the 2 replication checkpoints and your custom created ones.
To determine, which of the 2 replication checkpoints is which mountpoint is a bit tricky. But as they should not be used by anything else than the ip-replication itself, it shouldn´t matter which one is which one.
matteo_marchett
112 Posts
0
May 27th, 2015 06:00
Thanks Peter it makes sense.
I've forced some new data to be copied to the secondary site and I've checked that the two additional snapshots are "rotating" as you said.
drwx-----x 3 cheesep staff 1024 May 27 14:42 2015_05_27_14.43.13_Europe_London
drwx-----x 3 cheesep staff 1024 May 27 14:42 2015_05_27_14.42.33_Europe_London
drwx-----x 3 cheesep staff 1024 May 27 14:45 2015_05_27_14.45.53_Europe_London
drwx-----x 3 cheesep staff 1024 May 27 14:42 2015_05_27_14.43.13_Europe_London