Restoring deleted files from CTA recycle log

General Discussion
1

Hi,

I am trying to use/test cta-restore-deleted-files tool following

I have done

eos vid set map -grpc key:UUID vuid:2 vgid:2

and

eos -r 0 0 vid add gateway 130.246.182.56 grpc

where the IP is of the admin frontend on which I have created /etc/cta/eos.grpc.keytab

eosantarespreprodfac antares-eos98.scd.rl.ac.uk:50052 UUID

and defined

cta.ns.config /etc/cta/eos.grpc.keytab in /etc/cta/cta-frontend-xrootd.conf

but I get an error

[root@cta-front03 ~]# cta-restore-deleted-files --id 4295383767 --instance eosantarespreprodfac --fid 200507323 --filename /eos/antarespreprodfac/factest/random_22102025_8G_3 --vid JL5698 --copynb 1 -d
Unable to open file /eos/antarespreprodfac/factest/random_22102025_8G_3
[root@cta-front03 ~]#

What am I missing?

Many thanks,

George

2

Hello George,
I had to recently go through the same instruction from the same handson reference documentation.

A few notes about the configuration I could read in your message:

  1. are you certain that the eos grpc gateway is running on port 50052? On my side I configured gRPC gateway on the same 50051 port as the handson instruction and was fine
  2. you should add IPv6 of the admin frontend on eos mgm between square brakets: eos -r 0 0 vid add gateway [admin frontend IPv6] grpc usually IPv6 is preferred over IPv4 those days (it was not the case in CI 2 years ago).

Then regarding the restoration process itself you should restore on your admin frontend using a restore invariant between eos and CTA: archive id is the only thing you need to uniquely identify the file on both sides of the coin and you also need to tell the frontend which disk instance it should contact via gRPC to restore the file. Indeed CTA should offer the possibility to restore a file to another disk instance.
Therefore this should be sufficient: cta-restore-deleted-files --id 4295383767 --instance eosantarespreprodfac

Do not forget that the restored file fileid will be different from the deleted file fileid.

In terms of eos and CTA versions I used for my last restore 2 weeks ago: EOS: 5.3.38-1 and CTA frontend version: 5.11.18.3-1 (latest stable).

It should be OK with previous versions as this code should not have changed during the last 12 months.

Best regards,
Julien Leduc

3

Hi Julien,

Thanks for the reply.

Actually, I am confused about which gRPC port to use. I started with 50051 as per your documentation but then I noticed the following the MGM log

260915 11:59:52 time=1789469992.856090 func=RunWnc level=INFO logid=static.............................. unit=mgm@antares-eos98.scd.rl.ac.uk:1094 tid=00007feb11ffb640 source=GrpcWncServer:321 tident= sec=(null) uid=0 gid=0 name=- geo="" xt="" ob="" gRPC server for EOS-wnc is running on port 50052.

I will change add IPv6 in the eos add gateway command but our admin frontend is a single stack vmware host. In general, only our Tier-1 EOS nodes are dual stack while all our CTA nodes are not only IPv4 but they are also in a subnet with private IP addresses.

For the record I am running EOS 5.3.23 and CTA 5.11.14.

Best,

George