SQL backups fail only on Sunday

Using NW 7.6.1 build 472.

I have a group with two SQL servers on a schedule of full backups every day but Sunday. On Sunday, the schedule in the group/properties/advanced tab is set to "skip". Each of the clients in that group use the same schedule. Yet, every Sunday the job runs anyway but the backups fail. When I return on Monday and rerun the job, they backup just fine. Wondering if, rather than skipping the job on Sunday, NW actually tries to do a backup, then realizes it does not need to do anything, as it is a Sunday, so it then aborts the job and they show up in the log as a failure for each of the DBs on the client.

A typical error log is shown below:

29085 1324277134 5 0 0 6188 6988 0 chyws0016.us.qdx.com (pid6988) 36 Microsoft SQL Server Provider error: 0
38006 1324277134 5 0 0 6188 6988 0 chyws0016.us.qdx.com (pid6988) 3 %s. 1 24 161 Write on "Legato#55ef8f73-e5ce-464e-8bf2-9afdf87ea36e" failed: 995(The I/O operation has been aborted because of either a thread exit or an application request.)
38006 1324277134 5 0 0 6188 6988 0 chyws0016.us.qdx.com (pid6988) 3 %s. 1 24 190 A nonrecoverable I/O error occurred on file "Legato#55ef8f73-e5ce-464e-8bf2-9afdf87ea36e:" 995(The I/O operation has been aborted because of either a thread exit or an application request.).
38006 1324277134 5 0 0 6188 6988 0 chyws0016.us.qdx.com (pid6988) 3 %s. 1 24 42 BACKUP DATABASE is terminating abnormally.

My question is why would there be a connection in the first place, if the schedule is set to "skip"? The log message makes it sound like something totally different. Of course, just to complicate the question, I have another SQL backup group using the very same schedule, that properly skips the Sunday work.The only clue here is that the group itself does not have any schedule override set, the clients in the group have the schedule. In the group that aborts, both the group and each client have the same schedule setting. Is the override not working properly because both the group and the clients have the same schedule settings?

Responses(2)

ble1

2 Intern

•

14.3K Posts

0

December 19th, 2011 14:00

When I read title of this post first thought was "That's easy, do not run backups on Sunday"

On more serious side, I have similar setup (except I do not do DB backups on Saturday and Sunday). The only difference is I have single client per group, but that should not matter. Schedule can be tricky sometimes and certain things may get easily overlooked so I would start from there. My first question is where do you setup schedule? OK, you already chedked that... but on top of that, on group level you may set level of the backup (it is either level or schedule - can't have both). If level is set group will override client setting meaning group level will bypass skip schedule if set on client. That would be for example reason why it is executed. Even so, it should not fail unless on Sunday you have something going on with DB which may impact success of backup.

In my env, I mostly use NMSQL 5.2.1 build 358, but recently I got one SQL hotel box where databases were a bit bigger. For no reason, I also started to get some failures similar to yours and I also noticed some strange PiT restore thingies. While I had no explanation for backup part, restore part looked like some issue with full and transaction log backup mess (I run those every 1h). To my surprise after reading release notes for 5.2.2 I found that this version is the first one which addresses concurent backups of DB and transaction logs (surprise surprise). For some reason too, my backup problems went away too (I did also set NSR_KEEPALIVE_WAIT though it didn't help with 5.2.1 and the only reason I did that was because manual backup from client worked fine for DB backups).

In summary, I would check if level on group is set and perhaps if you do not run 5.2.2 already to use that version. Last thing would be of course software bug - you can easily see that if you create additional group and place one client in one group and the other one in other - if they still run on Sunday then errrr... strange. If not, then you know that for some reason (bug) 2 SQL backups in single group do not play nice with skip schedule and this should be reported to support to be fixed.

david_knapp

1 Rookie

•

122 Posts

0

December 20th, 2011 06:00

Both servers with the problem were using version 5.2.1 of the SQL client, so I will plan to update those. The problem group did not have a level set, just a schedule. These DBs are small enough and the usage is such that there is no point in doing transaction backups, all of them are full. I have another group of SQL backups that did skip correctly, so don’t think it has something to do with multiple SQL skips, although that also crossed my mind. Thanks very much for the reply. I have had more trouble with SQL backups than all other types combined, but I also have little training in this product and am feeling around in the dark at times, so the problems may be caused by my own inexperience.

David Knapp

Please think of resource conservation before you print this message.

From: Hrvoje Crvelin

Sent: Monday, December 19, 2011 5:13 PM

To: Knapp, David W

Subject: New message: "SQL backups fail only on Sunday"

EMC Community Network

Re: SQL backups fail only on Sunday

created by Hrvoje Crvelin in NetWorker Support Forum - View the full discussion

View All

No Events found!