Hi All,



I'm monkeying around trying to a SANs data snapshots and Ingres backups
to agree with each other about what constitutes a backup and how to
recover it.



OS is Red Hat Enterprise Linux Server release 5.3 (Tikanga)



I've nailed the checkpoint is really a data snapshot part...that was
easy.



Now I'm trying to recover the snapshot....using rollforwarddb



To restore a volume from a snapshot, follow these steps:

1. Disconnect any iSCSI initiators from the volume. Follow the
instructions for your

operating system and initiator.



Now I know my test database is on SAN volume called 'bowtest'. Using the
SAN command 'volume select bowtest show' I can see that volume 'bowtest'
is tied to iSCSI Name:


iqn.2001-05.com.equallogic:0-8a0906-c0f132e04-3710000000c4a530-vol1.



I've set up ingres to have sudo access to the iscsiadm command.



I cant see a disconnect command in iscsiadm so I assume I have to log it
out. Is that the case?



If so it appears to me the command required is:

sudo iscsiadm --mode node \

--targetname
iqn.2001-05.com.equallogic:0-8a0906-c0f132e04-3710000000c4a530-vol1 \

--portal 192.168.0.210 \

--logout

But this gives output:

iscsiadm: initiator reported error (2 - not found)



If I do: sudo iscsiadm --mode discovery --type sendtargets --portal
192.168.0.210

Then I get response:

192.168.0.210:3260,1
iqn.2001-05.com.equallogic:0-8a0906-c0f132e04-3710000000c4a530-vol1



So I'm pretty sure I've got the correct target and portal.



Anyone got any idea what I'm doing wrong?



Marty

Hi All,



I've hacked around further...I really have no idea what I'm doing...and
have found that the --logoutall=all option appears to work.



I then do all the otherstuff to restore the snapshot. It looked cool.



I then try to reconnect the iSCSI initator with the -loginall=ALL
option. Again it looks cool. But the data area didn't mount.



ls -l /san

ls: reading directory /san: Input/output error



So I cant connect to the database as its data location under /san is
unreachable.



Ideas?



Marty





From: info-ingres-bounces (AT) kettleriver...ting (DOT) com
[mailto:info-ingres-bounces (AT) kettleriverconsulting (DOT) com] On Behalf Of
Martin Bowes
Sent: 10 July 2009 09:50
To: Ingres and related product discussion forum
Subject: [Info-Ingres] Dell Equalogic and ingres backups.



Hi All,



I'm monkeying around trying to a SANs data snapshots and Ingres backups
to agree with each other about what constitutes a backup and how to
recover it.



OS is Red Hat Enterprise Linux Server release 5.3 (Tikanga)



I've nailed the checkpoint is really a data snapshot part...that was
easy.



Now I'm trying to recover the snapshot....using rollforwarddb



To restore a volume from a snapshot, follow these steps:

1. Disconnect any iSCSI initiators from the volume. Follow the
instructions for your

operating system and initiator.



Now I know my test database is on SAN volume called 'bowtest'. Using the
SAN command 'volume select bowtest show' I can see that volume 'bowtest'
is tied to iSCSI Name:


iqn.2001-05.com.equallogic:0-8a0906-c0f132e04-3710000000c4a530-vol1.



I've set up ingres to have sudo access to the iscsiadm command.



I cant see a disconnect command in iscsiadm so I assume I have to log it
out. Is that the case?



If so it appears to me the command required is:

sudo iscsiadm --mode node \

--targetname
iqn.2001-05.com.equallogic:0-8a0906-c0f132e04-3710000000c4a530-vol1 \

--portal 192.168.0.210 \

--logout

But this gives output:

iscsiadm: initiator reported error (2 - not found)



If I do: sudo iscsiadm --mode discovery --type sendtargets --portal
192.168.0.210

Then I get response:

192.168.0.210:3260,1
iqn.2001-05.com.equallogic:0-8a0906-c0f132e04-3710000000c4a530-vol1



So I'm pretty sure I've got the correct target and portal.



Anyone got any idea what I'm doing wrong?



Marty

Hi All,

Further hacking has shown that all I have to do is mount /san!

Thank God for the sudo command.

And now it works! It lives, it breathes, it quacks like a duck! I think I'll call it Bernard.

It checkpoints in seconds and rollforwarddb +j pretty damn snappily gets stuck into doing journal recoveries.

I'll work on getting a real sized database over to the test box this coming week and get some timings and really put it through some hurdles and see what happens.

After that we'll try some snapshot replication!

Marty

________________________________

From: info-ingres-bounces (AT) kettleriver...ting (DOT) com on behalf of Martin Bowes
Sent: Fri 10/07/2009 10:19
To: Ingres and related product discussion forum
Subject: Re: [Info-Ingres] Dell Equalogic and ingres backups.



Hi All,



I've hacked around further...I really have no idea what I'm doing...and have found that the --logoutall=all option appears to work.



I then do all the otherstuff to restore the snapshot. It looked cool.



I then try to reconnect the iSCSI initator with the -loginall=ALL option. Again it looks cool. But the data area didn't mount.



ls -l /san

ls: reading directory /san: Input/output error



So I cant connect to the database as its data location under /san is unreachable.



Ideas?



Marty





From: info-ingres-bounces (AT) kettleriver...ting (DOT) com [mailto:info-ingres-bounces (AT) kettleriverconsulting (DOT) com] On Behalf Of Martin Bowes
Sent: 10 July 2009 09:50
To: Ingres and related product discussion forum
Subject: [Info-Ingres] Dell Equalogic and ingres backups.



Hi All,



I'm monkeying around trying to a SANs data snapshots and Ingres backups to agree with each other about what constitutes a backup and how to recover it.



OS is Red Hat Enterprise Linux Server release 5.3 (Tikanga)



I've nailed the checkpoint is really a data snapshot part...that was easy.



Now I'm trying to recover the snapshot....using rollforwarddb



To restore a volume from a snapshot, follow these steps:

1. Disconnect any iSCSI initiators from the volume. Follow the instructions for your

operating system and initiator.



Now I know my test database is on SAN volume called 'bowtest'. Using the SAN command 'volume select bowtest show' I can see that volume 'bowtest' is tied to iSCSI Name:

iqn.2001-05.com.equallogic:0-8a0906-c0f132e04-3710000000c4a530-vol1.



I've set up ingres to have sudo access to the iscsiadm command.



I cant see a disconnect command in iscsiadm so I assume I have to log it out. Is that the case?



If so it appears to me the command required is:

sudo iscsiadm --mode node \

--targetname iqn.2001-05.com.equallogic:0-8a0906-c0f132e04-3710000000c4a530-vol1 \

--portal 192.168.0.210 \

--logout

But this gives output:

iscsiadm: initiator reported error (2 - not found)



If I do: sudo iscsiadm --mode discovery --type sendtargets --portal 192.168.0.210

Then I get response:

192.168.0.210:3260,1 iqn.2001-05.com.equallogic:0-8a0906-c0f132e04-3710000000c4a530-vol1



So I'm pretty sure I've got the correct target and portal.



Anyone got any idea what I'm doing wrong?



Marty