Shadow server issue

[Simon Watkins]

Hello, thanks in advance for your help                    

We have a server A as a Source Server and Server B as a Shadow Server. Both are identical Windows 

Server 2012 R2 Virtual boxes running Cache for Windows (x86-64) 2014.1.3 (Build 775U) on x64 

architecture.

On May 22nd, due to operational issues, we were supposed to migrate one of the servers 

(Server B) to new hardware in a brand new environment. We prepared everything on Tuesday afternoon 

for the migration and shut down Caché on Server B safe in the knowledge that the journal files 

would continue being created on Server A for 2 days as per the default settings.

However, as is often the case, there were problems with the migration for a number of different 

servers (not Caché ones), and our Data Centre team aborted the plan that afternoon, with a plan to 

resume the next day. I wasn’t too worried by this since our servers are set to keep the journals 

for what we thought were 2 days. Anyway, the migration was completed successfully this morning 

(24th May). 

I started up Caché on Shadow Server (B) this morning (24th May), expecting the journals from 

Tuesday afternoon (less than 2 days earlier) to be applied, and received these error messages;

#       Time    PID     $ZERROR Text    Number or reoccurrences         First Occurred At

1       2018-05-24 10:44:16    6520    <ZABOR>EXEC1+5^SHDWCBLK        ERROR #1029: Shadowing aborted on error        

        24/05/2018 10:44

2       2018-05-24 10:44:15    6428            ERROR #1029: Shadowing aborted on error               24/05/2018 10:44

3       2018-05-24 10:44:15    6428    <ZMISC>RCVB+31^SHDWCBLK        ERROR #1033: Requested journal file 

'<path> 20180522.003' does not exist on the source                24/05/2018 10:44

To our horror we realised that the setting to keep journals for 2 days, was 2 calendar days, not 

rolling days, so our journals from 22nd May afternoon were lost.

Now, I am able to restore these journal files from Source Server A from yesterday’s routine 

backups. This date’s backup should have all the journal files on the Source Server A from 

22-May-2018 since the backup was created on the 23rd May at 20:47 and 22.58.

The question I’ve been building up to is; can I just copy the missing but now restored journal 

file(s) into the Source Server A directory (might as well restore all 

20180522.00x files), resume journaling on Server B, and it’ll automatically recover, or do I 

need to do anything else?

[Andrew Makinson]

it might be easier to stop cache on the primary server databases and copy the cache.dat files over and create new mirror files. And then restart the Mirrors.

There is a step by step Guide to doing this in the intersystems documents 

Sent from my iPhone

--
--
IRIS, Caché, Ensemble
---
You received this message because you are subscribed to the Google Groups "IRIS, Caché, Ensemble" group.
To unsubscribe from this group and stop receiving emails from it, send an email to intersystems-publi...@googlegroups.com.
To post to this group, send email to intersystems...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Lyle Marti]

Once you restore the journal files to the source, you will overcome the error on server B.  The requested journal will exist and will be pulled over to server b. The other thing you can do is just refresh server B from server A, unless hardware changes prevent it.  If you do that then you know you are in sync.  

Lyle

--

--
IRIS, Caché, Ensemble
---
You received this message because you are subscribed to the Google Groups "IRIS, Caché, Ensemble" group.

To unsubscribe from this group and stop receiving emails from it, send an email to intersystems-public-cache+unsub...@googlegroups.com.
To post to this group, send email to intersystems-public-cache@googlegroups.com.

[billr...@gmail.com]

I believe the answer is yes.  Been a few years since I've had to do that though.

Bill

[Venkatesh Gadi]

if you could get some downtime on shadow server and your database is not too huge , In these cases it is always better to do the total refresh again from production and start shadowing.

[Simon Watkins]

To post to this group, send email to intersystems...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

Just to let you know, this is exactly what we did. Thank you for your help.