The save set DISASTER_RECOVERY:\ should be ran by itself, as this save set will be backing up all the critical volumes, and will be used to perform DR on servers running Windows 2008, Windows 2008 R2 and Windows 7.
You cannot run the backup for C:\ and DISASTER_RECOVERY:\ together nor at the same time in different groups.
ALL will not include DISASTER_RECOVERY:\ save set, so we would need to clarify this first.
Please take a look at the documentation (NetWorker Administrator's Guide 7.6.2
I have the same problem. I have been having vague problems with VSS which I am trying to solve. Upgrading 7.6.1.5 to 7.6.2.1 introduced the disaster recovery and the entire backup process is different it seems. No clue what OSSR_ means. Still reading up ;-).
As for the disaster recovery not being part of 'all' and should not be run with other backups. Please look at article ESG121662 'What is the difference between Online and Offline Disaster Recovery in NetWorker 7.6.2'. It specifies that disaster_recovery:\ is part of the 'all' and is run with all others.
On the original item (I have had the same problem): 2 things I have come accross
1- apperantly in some cases VSS backups cannot be done at the same time so imho parellism should be set at 1 (!). There is a microsoft article specifying this with an accompanying hotfix of you would like to fix this (http://support.microsoft.com/kb/975928)
2- I have had problems with servers which are separated by a Cisco ASA firewall (5540). I also happen to be the network engineer and I knew that there was a setting that would remove a tcp session if no more network traffic was seen for one hour (!). So I set this setting to keep the tcp session going forever. and this seemed to work as well. I found that the backup would end exactly 1 hour after it started so this triggered me to search for this item. The setting I am talking about is located under firewall. service policy rules. global-class (or the class you are using). Then the third tab 'rule actions' and in the sub tabs 'connection settings' => connection timeout. Default is 1 hour and send reset to tcp endpoints is off. I set this last one to on as a TCP reset has the same effect on endpoints as a TCP reconnect if connection is lost or incorrect for some reason. Maybe something like this may be the case with you too.
Carlos, I believe you are mistaken. The All saveset does indeed include DISASTER_RECOVERY and this is mentioned a number of times in the Admin Guide and other documentation. To not include it would need an explicit "skip". However, that means we now need to block out time for additional disaster recovery backups in what are already too short backup time windows. This probably would be disruptive to production. The system should just understand not to try to VSS snaphot the same drive more than once at the same time. Hopefully this has been or will be addressed in a 762 hotfix. I need to review the fixes to be sure. The review alone is a tedious task.
Maarten, the OSSR_ save sets are all explained starting on page 600 of the admin guide. I'm reluctant to apply the MS hotfix, since the error it is claimed to repair is not the error I'm receiving. Most firewall session errors can be aleviated by setting a system variable NSR_KEEPALIVE_WAIT on the client. it is the number of seconds between session pings to keep the connection open while the client is working. We use 28 and it works well. search for it in the release notes for details
Solving my own problems with Windows 2008 and networker 7.6.2 I found a problem not with networker but with Microsoft Windows. Apperantly there is a problem with windows 2008 and some update (it doesn't appear on 2008R2). There is a work around but it has to be re-applied every time patch tuesday has come. It has to do with the VSS system writer which disappeared.
It is mentioned in http://support.microsoft.com/kb/2009272 and on EMC powerlink. However this is a temporary measure because after each windows update you have to re-do the fix. And the fix has to do with resetting permissions on the winsxs folder (which is not something I am keen to do, being the windows core and all). After a permissions reset the system writer re-appears when you look at the 'vssadmin list writers' command
Maybe this will help you as well. Currently I am thinking of starting a Microsoft call on this issue to get it fixed permanently. But I am pressed for time to actually do so.
36115_carloscor
217 Posts
0
August 19th, 2011 04:00
Hello,
I think there is something unclear here.
The save set DISASTER_RECOVERY:\ should be ran by itself, as this save set will be backing up all the critical volumes, and will be used to perform DR on servers running Windows 2008, Windows 2008 R2 and Windows 7.
You cannot run the backup for C:\ and DISASTER_RECOVERY:\ together nor at the same time in different groups.
ALL will not include DISASTER_RECOVERY:\ save set, so we would need to clarify this first.
Please take a look at the documentation (NetWorker Administrator's Guide 7.6.2
Attached screenshot.
Thank you.
Carlos.
1 Attachment
DR.JPG
maarten3
2 Posts
0
August 24th, 2011 05:00
I have the same problem. I have been having vague problems with VSS which I am trying to solve. Upgrading 7.6.1.5 to 7.6.2.1 introduced the disaster recovery and the entire backup process is different it seems. No clue what OSSR_ means. Still reading up ;-).
As for the disaster recovery not being part of 'all' and should not be run with other backups. Please look at article ESG121662 'What is the difference between Online and Offline Disaster Recovery in NetWorker 7.6.2'. It specifies that disaster_recovery:\ is part of the 'all' and is run with all others.
On the original item (I have had the same problem): 2 things I have come accross
1- apperantly in some cases VSS backups cannot be done at the same time so imho parellism should be set at 1 (!). There is a microsoft article specifying this with an accompanying hotfix of you would like to fix this (http://support.microsoft.com/kb/975928)
2- I have had problems with servers which are separated by a Cisco ASA firewall (5540). I also happen to be the network engineer and I knew that there was a setting that would remove a tcp session if no more network traffic was seen for one hour (!). So I set this setting to keep the tcp session going forever. and this seemed to work as well. I found that the backup would end exactly 1 hour after it started so this triggered me to search for this item. The setting I am talking about is located under firewall. service policy rules. global-class (or the class you are using). Then the third tab 'rule actions' and in the sub tabs 'connection settings' => connection timeout. Default is 1 hour and send reset to tcp endpoints is off. I set this last one to on as a TCP reset has the same effect on endpoints as a TCP reconnect if connection is lost or incorrect for some reason. Maybe something like this may be the case with you too.
regards,
Maarten
pdumigan
40 Posts
0
October 21st, 2011 05:00
Carlos, I believe you are mistaken. The All saveset does indeed include DISASTER_RECOVERY and this is mentioned a number of times in the Admin Guide and other documentation. To not include it would need an explicit "skip". However, that means we now need to block out time for additional disaster recovery backups in what are already too short backup time windows. This probably would be disruptive to production. The system should just understand not to try to VSS snaphot the same drive more than once at the same time. Hopefully this has been or will be addressed in a 762 hotfix. I need to review the fixes to be sure. The review alone is a tedious task.
Maarten, the OSSR_ save sets are all explained starting on page 600 of the admin guide. I'm reluctant to apply the MS hotfix, since the error it is claimed to repair is not the error I'm receiving. Most firewall session errors can be aleviated by setting a system variable NSR_KEEPALIVE_WAIT on the client. it is the number of seconds between session pings to keep the connection open while the client is working. We use 28 and it works well. search for it in the release notes for details
maarten3
2 Posts
0
October 24th, 2011 00:00
Solving my own problems with Windows 2008 and networker 7.6.2 I found a problem not with networker but with Microsoft Windows. Apperantly there is a problem with windows 2008 and some update (it doesn't appear on 2008R2). There is a work around but it has to be re-applied every time patch tuesday has come. It has to do with the VSS system writer which disappeared.
It is mentioned in http://support.microsoft.com/kb/2009272 and on EMC powerlink. However this is a temporary measure because after each windows update you have to re-do the fix. And the fix has to do with resetting permissions on the winsxs folder (which is not something I am keen to do, being the windows core and all). After a permissions reset the system writer re-appears when you look at the 'vssadmin list writers' command
Maybe this will help you as well. Currently I am thinking of starting a Microsoft call on this issue to get it fixed permanently. But I am pressed for time to actually do so.
Regards.