Implementing a Backup/Replication Servers further reduces downtime, with the previous night data positioned followed by hourly snap shots and transaction logs written to Backup server.
Assume Main server is dead !!!! Restore the transaction log on Backup Server and switch users over..... Attend to dead Server, at the end of the day position data back onto main server, reset users and a way we go. I have found with some other databases, you can get caught out when you do a restore and oops !!!! cant restore. So I take the added precaution (where applicable) of shutting down the services and taking a copy of the physical database as a pre-backup process. In additional to database maintenance. To prevent a potential 24 hour loss of data in all cases - A snap shot of the database along side other data and placed on a alternating/rotating 2.5" USB HDD ready for the last doctor to take off-site as they leave. I can't bear the thought of doctor/staff re-entering a days worth of data. Jerry Karpluk GPHelpDesk.net -----Original Message----- From: Barry Lollo [mailto:[EMAIL PROTECTED] Sent: Thursday, 24 November 2005 2:20 PM To: [EMAIL PROTECTED]; General Practice Computing Group Talk Subject: RE: [GPCG_TALK] A techies view of PractiX Andrew, <quote> Sbsbackup snapshots mysql perfectly to dds72. Praticx is good that all db's are now simple, unlike plexus where one is not and snapshot fails. </quote> I suspect this is "Simple Recovery Model". Am I right in assuming that !! While I appreciate the lack of Management required to maintain a DB in simple recovery mode, is this the best model for a PMS ?? I'm asking in terms of the potential to loose data. As SQL is transactional, a simple recovery model uses "Circular Logging", meaning that the log file remains a Specific size, and transactions are overwritten in a FIFO strategy. Therefore, ALL transactions since the last full backup, cannot be relied upon to be in the log file, if a failure occurs and recovery is warranted. Lets say that worst case scenario occurs, and the system has a catastrophic crash in the database for some reason, at 4pm in the afternoon. What is the recovery plan ?? In a "Full recovery Model", the 1st step is to try and backup the Transaction Log, as it contains all the transactions since my last full backup (ie: all the work that occurred that day). Then I would restore my full backup from last night, and any other transaction log backups (if you did any), and then restore the transaction log I backed up before I touched anything. Finally telling SQL backup to "Recover" the Database, effectively getting the database back to a level of integrity to the "point of Failure", or if desired a "Point in Time". I can see that the Simple recovery model, that you have deployed, would get me back to last night, and then we would have to re-enter all the data from the current day, which if we are heading down the path of computerisation, this data will be hard to get back IMHO. I have setup Maintenance Plans on most of my SQL servers, for Backups, and Optimisations. If you like I can give you some idea of how these happen and the Recovery Plans If that is of interest to you. Of course if I'm barking up the wrong tree, I'll take it all back.!!! Regards Barry Lollo Townsville Division. -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Andrew Cameron Sent: Thursday, 24 November 2005 8:10 AM To: 'General Practice Computing Group Talk' Subject: RE: [GPCG_TALK] A techies view of PractiX Thanks forthat info, Very timely seeing as my site is going live today or tomorrow ! I was reading your post... And was thinking...cap mssql mem usage, Then you did . 512mb on pc's seems a little low ? I don't use ir install exchange so i am 500 mb ahead of you on ram. Oh and i avoid having any path download software on the server too if i can aviod. Plexus services...running minimised in system tray on server console It will open another instance if you rdp in and then they both crash. Running rdp session as /console seems to help. I have asked mss to run it as a proper nt service, Not a pretend one like now. Painful if it crashes as no path is imported. Sbsbackup snapshots mysql perfectly to dds72. Praticx is good that all db's are now simple, unlike plexus where one is not and snapshot fails. So you have found yet another i.t. co that charges heaps and hides all the errors JOY. Wish we could shamne them on national tv. I always use 2 partitions c: for o/s and D; for data. Sometime z: for swap file. Good to share experiences re mss. Thanksfully i have a good relationship with rachel their local trainer, She lets me away with stuff and we always chat heaps. I sometimes get calls from her about other sites seeking advice and its worth giving her the time to build the Relationship. Fingers crossed, Andrew Cameron Andrew Cameron Technologies Gold Coast Australia [EMAIL PROTECTED] > -----Original Message----- > From: Jerry Karpluk [mailto:[EMAIL PROTECTED] > Sent: Wednesday, 23 November 2005 9:17 AM > To: [EMAIL PROTECTED]; 'General Practice Computing Group Talk' > Subject: RE: [GPCG_TALK] A techies view of PractiX > > Hi Andrew, > > I would like to share my experience with a site that we were recently > invited to review who upgraded to PractiX about 8 weeks ago. > > The systems has just been running like a dog, to the point of > un-usable (long delays) from the word go. Doctors and staff were > literally pulling their hair out. > > IBA performed a pre-check of the system to ensure it met minimum > requirements prior to upgrade and matches your current configuration. > > They started with an IBM Server 1 x 2.8Ghz XEON, WIN2003SBS, 2GB RAM, > RAID 5 (143GB - 15k rpm) and were recommended by the current IT > Support company to install an additional XEON CPU, up all the > workstation (21 off them) to 512MB, Gigabyte Switch a lot of $$$ spent > - same outcome. > > Database size roughly 7GB > > Pass the buck around began (Hardware <-> Software), without either > party willing to go the extra mile and collaboratively resolve the > problem in the best interest of the patient care and their client. > > We identified over 15 issues which potentially contributed to the > problem. IBA were provided the first opportunity to review and perform > post conversion activities after we identified 1% HDD space > left. Now with 70Gb free - no improvement. > > After resolving the other 14 remaining issues - slight improvement. > Even setting up another server - same outcome. This conclusively > showed that the issue must be with the databases themselves. Referred > the matter back to IBA were it was suggested that a performance > (index) check was performed on the database. > > 24 hours later, index were optimised and problem fixed !!!! > > At this stage, we still are suggesting that the Server could do with > an additional 1.0 - 2.0 Gb Ram as memory utilisation is at 90 - 95%. > > Looking at memory utilization : > > WINDOWS 2003 500Mb Ram > Exchange 540Mb > Plexus Classic 150Mb > MS-SQLSERVER Grabs whatever it can (to the max 1.0Gb) it > seems to grab > about 75-150Mb RAM per user. > > We repeatedly get mspservices.exe (amongst other apps) crashing due to > insufficient memory or it's a possible memory leak. It just seems that > MS-SQL does not release memory in a timely fashion. I'm unable to get > a definitive answer from IBA as to what memory PracticX ideally > requires for service MS-SQL with a 7Gb database + factoring growth > > MS-SQL offers memory management tools, so we have allocated a fixed > amount and that has resolved the problem. > > This site has been setup with a single Drive C: partition ....WINDOWS > 2003 SBS should be ideally setup with separate partitions : > > C: WIN2003 O/S > E: Exchange > P: Practice Management Software > S: Swap file > > This will ensure that databases files are less prone in becoming > fragmented, as MS-SQL grow automatically at a rate of 10% (can be > adjusted) > > The site has been managed by an IT Support company with monthly SLA in > place > - paying excessively top $$$, onsite once and apparently SLA are > performed remotely - log files tell me differently. Sad to say Alert > and Monitoring was disabled which revealed the true state of the > system once enabled : Failed backups, server re-starts (middle of the > night) with a dodgy UPS, applications crashing, you name it - it was > there. > > Hope live day all goes well.. > > Cheers, > > > Jerry Karpluk > > > > -----Original Message----- > From: Andrew Cameron [mailto:[EMAIL PROTECTED] > Sent: Friday, 18 November 2005 1:06 AM > To: 'General Practice Computing Group Talk' > Subject: [GPCG_TALK] A techies view of PractiX > > A quick blog. > > A client is converting to PractiX after 6-8 years running MSS Classic. > > So first we replaced the network, server, pc's and switch. Dell > (single xeon) server running win2003 SBS with 2 gig ram and SCSI Raid. > Dell Optiplex's with 1 gig ram, winXP and 17" TFT's. > After it was bedded in for a few months and the bugs ironed > out they scheduled PracitX to be installed. > (that way the vendor can't blame the network if it all goes > pear shaped or just runs like a dog) > > The install/training plan is spread over 2 weeks which is good. > Installation first, then train, train and train some more. Then next > week the big conversion. 7 gig of data. > > *Annoying thing number one; > why can't the workstation install be just like installing MS word or > any other off the shelf product ? Why do I have to give authenticated > users full rights to the local MSP folder on their pc. > X 16 pc's. ? (i have better things to do with my time) > > *Annoying thing number two; > Why does the OBDC details have to be entered manually, with the > obilatory typo's and errors ? Shit even HCN manages to do that whilst > installing PS3 !!! > > Good things: > Boy PractiX is fast :) > Hooray the updates for items numbers etc is done from within the admin > module ! It goes zips off to the web and downloads and then installs. > > Looks like the wks install stays inside its MSP directory and doesn't > splatter itself into the system dir. > > Hmmn there are still some features in Classic that aren't in PractiX. > > Still I'm not looking forward to go live day one little bit. > > Hopefully I will remember to blog a little bit every other day. > > Cheers, > > Andrew Cameron. > > > > > _______________________________________________ > Gpcg_talk mailing list > [email protected] > http://ozdocit.org/cgi-bin/mailman/listinfo/gpcg_talk > > > > __________ NOD32 1.1289 (20051116) Information __________ > > This message was checked by NOD32 antivirus system. > http://www.eset.com > > _______________________________________________ Gpcg_talk mailing list [email protected] http://ozdocit.org/cgi-bin/mailman/listinfo/gpcg_talk _______________________________________________ Gpcg_talk mailing list [email protected] http://ozdocit.org/cgi-bin/mailman/listinfo/gpcg_talk __________ NOD32 1.1302 (20051124) Information __________ This message was checked by NOD32 antivirus system. http://www.eset.com _______________________________________________ Gpcg_talk mailing list [email protected] http://ozdocit.org/cgi-bin/mailman/listinfo/gpcg_talk
