Start a Conversation

Unsolved

This post is more than 5 years old

4028

January 3rd, 2012 10:00

GeoDrive questions

Hello,

I'm trying out GeoDrive on a Linux client server and after getting everything installed and working I'm left with a couple of questions.

1.  When I store a file on my geodrive mount I end up with three copies, the original one lets say /root/file1, the one on the geodrive mount lets call it /geodrive1/file1 and one in cache which would be /var/local/geodrive/.../cache/file1.  Even after waiting a few hours, I don't see the file being stubbed, at least, it's still taking up the same amount of space. 

This is my geodrive's mount point's config:

The way I thought it would work is you copy a file to your geodrive mount and in 30 seconds that file would be uploaded to the Atmos server and the file in cache would be stubbed after 45 seconds.

2. I created a geodrive mount, stored a file on it, confirmed that the file was copied to the atmos server, then I removed my geodrive.  Next I checked the atmos server and the file was still there.  So, I thought I'd be able to access it again by making another geodrive drive with the same information.  But when i did that I no longer see file1.  I copied file2 to the new geodrive and it went to the storage, so now I have file1 and file2 on the storage, but I can only see file2 on the geodrive.  Shouldn't I be able to see anything i have in storage?  What if I want to upload/download the same files from different servers?  I'm sure I'm missing something here.

Anyway, if anyone can help with these questions I'd really appreciate it.  I think I'm just missing something in the configuration somewhere but I'm unsure where to look.

Thanks!

5 Practitioner

 • 

274.2K Posts

January 3rd, 2012 22:00

For question #1, how did you verify the disk space, by file size? In GeoDrive, when a file was stubbed, the file size will remain unchanged. While the disk space will be freed. You could verify it by du command.  One note is that, the thread which responsible for the stub will be started every 30 minutes by default. That means, some time you need wait at least 30 minuts + stub delay time for a file stubbed. It is for the peroformance consideration. Stub thread have impact on the system performance.

For question #2, To uninstall GeoDrive, you will need first unmount GeoDrive folder. After uninstall, all user data and configuration will be kept in case reinstall or upgrade.

So for you case, when you reinstall the GeoDrive, you don't need create GeoDrive folder again. After install succeed, you just mount the GeoDrive folder, it would works.

  geodrivecfg --mount

In your question, you said make another geodrive drive with same information, it is install GeoDrive on another client or reinstall GeoDrive on the same client.  Are they using the same cloud root directory ("/GeoDriveTest1/")? Could you tell me more details? We did not observe such issue before, so need check if we have any step different.

Thanks.

5 Practitioner

 • 

274.2K Posts

January 3rd, 2012 22:00

For question #1, how did you verify the disk space, by file size? In GeoDrive, when a file was stubbed, the file size will remain unchanged. While the disk space will be freed. You could verify it by du command.  One note is that, the thread which responsible for the stub will be started every 30 minutes by default. That means, some time you need wait at least 30 minuts + stub delay time for a file stubbed. It is for the peroformance consideration. Stub thread have impact on the system performance.

For question #2, To uninstall GeoDrive, you will need first unmount GeoDrive folder. After uninstall, all user data and configuration will be kept in case reinstall or upgrade.

So for you case, when you reinstall the GeoDrive, you don't need create GeoDrive folder again. After install succeed, you just mount the GeoDrive folder, it would works.

  geodrivecfg --mount

In your question, you said make another geodrive drive with same information, it is install GeoDrive on another client or reinstall GeoDrive on the same client.  Are they using the same cloud root directory ("/GeoDriveTest1/")? Could you tell me more details? We did not observe such issue before, so need check if we have any step different.

Thanks.

5 Posts

January 4th, 2012 08:00

What I did:

1. Installed geodrive to two freshly built 5.6 vms.

2. On server1 I ran this command:

geodrivecfg --add  --mount-point /geodrive1 --cloud-server 10.209.82.232:80 --subtenantid mySubTenantid --uid GeoDriveTest1 --shared-secret mySharedSecret --root /GeoDriveTest1/

Which gave me:

[root@server1]~# df -h

Filesystem            Size  Used Avail Use% Mounted on

/dev/mapper/vg_root-lv_root

                       14G  2.6G   11G  21% /

/dev/mapper/vg_root-lv_tmp

                      2.0G   68M  1.8G   4% /tmp

/dev/sda1             190M   17M  164M  10% /boot

tmpfs                  24G     0   24G   0% /dev/shm

fuse                   14G  2.6G   11G  21% /geodrive1

Next, I unmounted the /geodrive with the geodrivecfg tool and changed the timeout settings to match what I posted before, then remounted it with geodrivecfg.

3.  Next I copied the geodrive install file to /geodrive1.

4.  After waiting a few minutes I used the firefox plugin to verify that the file was copied to the cloud storage. It was there in a folder called GeoDriveTest1.

Questions:

1.  At this point, if I do a 'geodrivecfg --delete' on server1's /geodrive1 how do I get it back? What I was doing was repeating the 'geodrivecfg --add' command I had a few lines up, which gets me access to the same folder but I can not see the file I previously uploaded.

2.  What do I need to do on server2 to mount /geodrive1 and see the file that server1 uploaded?  If I use the same 'geodrivecfg --add' command I had used on server1, I get access to the same folder but I can not see the file I previously uploaded.

5 Practitioner

 • 

274.2K Posts

January 4th, 2012 18:00

for #1, If you delete a folder, then want to get it back, you could repeat geodrivecfg --add with same information. The files will be synced down in the new folder.

Your steps:

   1) create a folder with geodrivecfg --add

   2) upload a file to that folder, check file to cloud via Atmos firefox.

   3) delete the folder via geodrivecfg --del

   4) create the folder again with the same parameters as step 1.

   5) then enter the folder, you could not see the files in step 2.

We tried such scenario before, but haven't observed such issue.

Please take a look at /var/log/atmos-geodrive (If this file is missing, please check /var/log/messages), if there are some error messages. We could know more why you could not sync down the cloud files.

for #2, your steps is correct. Please help double check if the date time synced with cloud server? If time not synced, the GeoDrive started, but worked in disconnected mode. You could store new files to it but it could not connect cloud for updates.  The same as #1, you could check /var/log/atmos-geodrive for error message.

5 Posts

January 5th, 2012 07:00

I tried your steps as shown here, but it didn't work out:

[root@server1]~# ls -l /geodrive1/

total 4

-rw-r--r-- 1 root root 159319156 Jan  4 11:09 atmos-geodrive-installer-rhel-56-64-no-gui-1032.bin

[root@server1]~# geodrivecfg --delete /geodrive1

Are you sure you want to delete /geodrive1?  All configuration for the drive and unsynchronized data will be lost. (y/n)> y

Unmounting /geodrive1

Removing /geodrive1 from /etc/fstab

Removing configuration directory /var/local/geodrive/conf/EF6F2CB1-777D-4E98-9D91-ECE2C2BFE971

Removing cache directory /var/local/geodrive/cache/EF6F2CB1-777D-4E98-9D91-ECE2C2BFE971

Removing logs directory /var/local/geodrive/logs/EF6F2CB1-777D-4E98-9D91-ECE2C2BFE971

Removing mount point /geodrive1

[root@server1]~# geodrivecfg --add  --mount-point /geodrive1 --cloud-server 10.209.82.232:80 --subtenantid mySubTenantid --uid GeoDriveTest1 --shared-secret mySharedSecret --root /GeoDriveTest1/

Success to store EFD1B35A-D57E-4286-AC59-6CC14BDB6A3D:ccb366a6855f491c910961a948a73153:GeoDriveTest1:sharedkey

[root@server1]~# ls -l /geodrive1/

total 0

[root@server1]~# date

Thu Jan  5 10:14:55 EST 2012

[root@server1]~# date

Thu Jan  5 10:16:12 EST 2012

[root@server1]~# ls -l /geodrive1/

total 0

[root@server1]~# ls -l /geodrive1/

total 0

[root@server1]~# date

Thu Jan  5 10:22:57 EST 2012

[root@ph679591]~# date

Thu Jan  5 10:24:45 EST 2012

[root@server1]~# ls -l /geodrive1/

total 0

[root@server1]~# date

Thu Jan  5 10:28:05 EST 2012

[root@server1]~# ls -l /geodrive1/

total 0

I would think that as soon as someone mounts / adds a drive, the files should start becoming avail.

Even still as I write this:

[root@server1]~# date

Thu Jan  5 10:32:36 EST 2012

[root@server1]~# ls -l /geodrive1/

total 0

I checked using the browser plugin and the file is still on the storage. 

I don't have a /var/log/atmos-geodrive.  I checked /var/log/messages and this is the only thing logged all day:

Jan  5 04:02:01 server1 anacron[16871]: Updated timestamp for job `cron.daily' to 2012-01-05

5 Practitioner

 • 

274.2K Posts

January 5th, 2012 20:00

Looks like file sync down never works on your testbed. What's the version number of your Atmos system?

When you install 2 GeoDrive clients with same information, if you create filea in GeoDrive client #1, could you see it in GeoDrive client #2?

I suppose the issue is not because of GeoDrive delete.

For logging,

please check /etc/init.d/rsyslog status.  If it is stopped, please start it first. Make sure it is running, then logging will be recorded. Thanks.

5 Posts

January 6th, 2012 11:00

The atmos version is 1.4.1a.

I created a /geodrive2 with the same information.  I put a file called filea on /geodrive1 and confirmed it was there using the browser plugin.

However, I can not see it from /geodrive2:

[root@server1]~# ls -la /geodrive1

total 28

drwxr-xr-x  3 root root 4096 Jan  6 14:19 .

drwxr-xr-x 26 root root 4096 Jan  6 14:17 ..

-rw-r--r--  1 root root   24 Jan  6 14:19 filea

drwxr-xr-x  2 root root 4096 Jan  5 10:14 .geodrive

[root@server1]~# ls -la /geodrive2

total 20

drwxr-xr-x  3 root root 4096 Jan  6 14:17 .

drwxr-xr-x 26 root root 4096 Jan  6 14:17 ..

drwxr-xr-x  2 root root 4096 Jan  6 14:17 .geodrive

5 Practitioner

 • 

274.2K Posts

January 10th, 2012 18:00

I reply this via email last week, but looks like it was not posted here by the system. So I paste again:

Could you execute this command on your cloud server?  It is the node that cloud server specified

rest_client.rb -f readobject -p / -k -c /GeoDriveTest1/ --includemeta --systemtags size --usertags geodrive.common.encoding -r /tmp/1

The output will be recorded in /tmp/1.

Will it looks like following one?

       

       

                4d365acda120b59104d365b8325f8804d940a51b34e8

                regular

                file1

               

                       

                                size

                                29

                       

               

               

               

       

       

Please help get rsyslog work, so we could get error message in /var/log/atmos-geodrive (start rsyslog service, it might require stop syslog service). It could speed up our investigation.

Thanks.

No Events found!

Top