Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore.
All, Does anyone have experience of how long OWLIM should take to go through it's transaction log and update all indicies? We have loaded a large volume of data (~7.5 GB of ttl files) into a repository and have a backup file of 8.7GB. We stopped Tomcat and on restart it is now going through transactions, presumably synching any differences between the disk and memory indicies when shutdown. Many thanks, Nic From: owlim-discussion-boun...@ontotext.com [owlim-discussion-boun...@ontotext.com] on behalf of Barry Bishop [barry.bis...@ontotext.com] Sent: 01 April 2012 09:12 To: Barry Norton Cc: owlim-discussion@ontotext.com Subject: Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. Hello all, This message is displayed (and OWLIM refuses to continue) only after the following sequence of events: 1. OWLIM is stopped and not shutdown properly 2. OWLIM is restarted and the recover process starts 3. OWLIM is stopped again without allowing the recovery process to finish 4. OWLIM is restarted At this point, the recovery code is aware of the backed up indices, some half built replacement indices and the transaction log. Without knowing which set of indices is correct, it informs the user and stops. (This is not an ideal situation and it is one that can not occur in OWLIM 5) The important point is that you should always allow OWLIM time to shutdown properly and should not kill tomcat, even if it seems to take a long time. If you have got in to this situation, the probable means to get started again is to shut down tomcat and delete any *.old files in the owlim storage folder. Then restart and allow time for OWLIM to replay its transaction log. The cause of this is that OWLIM 4 (and before) uses a lazy writing strategy for dirty pages. These are only swapped to disk when memory has been exhausted and more pages are required. In other words, the state of the indices on disk and in memory can diverge quite considerably and in the event of a sudden stop, the correct indices must be rebuilt using the transaction log (a file called 'backup' in the OWLIM storage folder). OWLIM 5 still has this mode of operation available (it is called 'fast' mode), but the default mode flushes updates to disk at the point of each commit. It does this in such a way, that a failure at any time leaves the disk image consistent with the state of the database at the end of the last successful commit operation. For sustained loading operations, this can have lower performance, but the advantages are: 1. instant shutdown time 2. instant recovery time (even if the last instance was stopped abruptly) 3. improved concurrency of reads and write operations This last point means that export the database programmatically is now a viable means to backup the database. To summarise: OWLIM 4 - this error occurs during initialisation when the OWLIM process has been killed, restarted and not allowed time to recover properly, i.e. killed again. It can be avoided by giving owlim enough time to shutdown properly (which can be a long time after loading large amounts of data on a machine with large memory). If it does happen, shutdown, delete the *.old files, restart and wait. OWLIM 5 - using the default transaction mode, this problem will not occur. The fast transaction mode is available when administrators need maximum performance and reduced query/update concurrency, but will have no guarantees about data security if OWLIM is killed before shutting properly - switching between transaction modes requires a restart. I hope this helps, barry Barry Bishop OWLIM Product Manager Ontotext AD Tel: +43 650 2000 237 email: barry.bis...@ontotext.com www.ontotext.com On 01/04/12 01:29, Barry Norton wrote: Guys, how are you shutting down your database that it needs to be recovered so often? And are you interrupting the recovery process? Barry On 31/03/2012 20:44, Blasetti, Luciano (CIOK) wrote: Hi All, I also experienced quite frequently this issue on a tomcat instance running only openrdf-sesame (OWLIM 4.3 SE) and openrdf-workbench. No way to restore the database. Moreover, since there's no out-of-the-box functionalities to perform a hot backup on a non clustered configuration (and the programmatic export can't work on large triplestore due to timeout issues), this strange behavior leads to data loss. I hope that the forthcoming OWLIM 5.0 will be more robust and reliable. Regards, Luciano From: owlim-discussion-boun...@ontotext.com [owlim-discussion-boun...@ontotext.com] On Behalf Of Marek Šurek [marek_su...@yahoo.co.uk] Sent: Saturday, March 31, 2012 11:18 AM To: Monika Solanki; owlim-discussion@ontotext.com Subject: Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. Hi Monika
Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore.
Thanks for the quick response Barry. We figured that we would have to wait for it to complete. I have a project manager breathing down my neck for an ETA on this so if you are able to give us an idea of how long 'a long time' could be that would be helpful. The transaction file is up to around 9000 after 1 hour 40 mins and as stated before the backup is 8.7 GB. Is it possible to estimate the number of transactions in the backup file? Thanks again for your help. Nic From: Barry Bishop [barry.bis...@ontotext.com] Sent: 13 April 2012 12:33 To: GOULD, Nicholas Cc: Barry Norton; owlim-discussion@ontotext.com Subject: Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. Hello Nicholas, This can take a long time, I'm afraid. If the indices are damaged then OWLIM will replay it's entire transaction log in order to recreate them. My suggestion is just to wait until it completes. Terminating the recovery process means that the next restart will fail to initialise and cause the the message in the subject of this email to be displayed. The good news is that the transaction mechanisms in OWLIM 5.0 have been completely refactored in order to fix these kinds of inconveniences. We will release this version next week. Please see: http://owlim.ontotext.com/display/OWLIMv50/OWLIM-SE+Indexing+Specifics#OWLIM-SEIndexingSpecifics-TransactionMode I hope this helps, barry Barry Bishop OWLIM Product Manager Ontotext AD Tel: +43 650 2000 237 email: barry.bis...@ontotext.com www.ontotext.com On 13/04/12 13:24, GOULD, Nicholas wrote: All, Does anyone have experience of how long OWLIM should take to go through it's transaction log and update all indicies? We have loaded a large volume of data (~7.5 GB of ttl files) into a repository and have a backup file of 8.7GB. We stopped Tomcat and on restart it is now going through transactions, presumably synching any differences between the disk and memory indicies when shutdown. Many thanks, Nic From: owlim-discussion-boun...@ontotext.com [owlim-discussion-boun...@ontotext.com] on behalf of Barry Bishop [barry.bis...@ontotext.com] Sent: 01 April 2012 09:12 To: Barry Norton Cc: owlim-discussion@ontotext.com Subject: Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. Hello all, This message is displayed (and OWLIM refuses to continue) only after the following sequence of events: 1. OWLIM is stopped and not shutdown properly 2. OWLIM is restarted and the recover process starts 3. OWLIM is stopped again without allowing the recovery process to finish 4. OWLIM is restarted At this point, the recovery code is aware of the backed up indices, some half built replacement indices and the transaction log. Without knowing which set of indices is correct, it informs the user and stops. (This is not an ideal situation and it is one that can not occur in OWLIM 5) The important point is that you should always allow OWLIM time to shutdown properly and should not kill tomcat, even if it seems to take a long time. If you have got in to this situation, the probable means to get started again is to shut down tomcat and delete any *.old files in the owlim storage folder. Then restart and allow time for OWLIM to replay its transaction log. The cause of this is that OWLIM 4 (and before) uses a lazy writing strategy for dirty pages. These are only swapped to disk when memory has been exhausted and more pages are required. In other words, the state of the indices on disk and in memory can diverge quite considerably and in the event of a sudden stop, the correct indices must be rebuilt using the transaction log (a file called 'backup' in the OWLIM storage folder). OWLIM 5 still has this mode of operation available (it is called 'fast' mode), but the default mode flushes updates to disk at the point of each commit. It does this in such a way, that a failure at any time leaves the disk image consistent with the state of the database at the end of the last successful commit operation. For sustained loading operations, this can have lower performance, but the advantages are: 1. instant shutdown time 2. instant recovery time (even if the last instance was stopped abruptly) 3. improved concurrency of reads and write operations This last point means that export the database programmatically is now a viable means to backup the database. To summarise: OWLIM 4 - this error occurs during initialisation when the OWLIM process has been killed, restarted and not allowed time to recover properly, i.e. killed again. It can be avoided by giving owlim enough time to shutdown properly (which can be a long time after loading large amounts of data on a machine with large memory). If it does happen, shutdown
Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore.
Hi Nicholas, Yikes, sorry to hear that you are in such a tight spot. Do you know roughly how big your transactions are? This would provide a way to make a guess about how many are in the backup file. I just tried an experiment: LUBM(1) is made up of 15 files of approximately 10,000 statements each. The size of the backup file is 4MB. If you (by some fluke) have the same average transaction size then I would guess there are this many transactions: (8000MB/4MB)*15 = 30,000 transactions. This is very 'back of a fag packet stuff' and depends completely on your average transaction size. If it was 5000 statements then expect 60,000 transactions, if it was 30,000 statements per transaction then expect about 10,000 transactions. Sorry, but this is the best I can do at this stage. Best of luck, barry On 13/04/12 13:47, GOULD, Nicholas wrote: Thanks for the quick response Barry. We figured that we would have to wait for it to complete. I have a project manager breathing down my neck for an ETA on this so if you are able to give us an idea of how long 'a long time' could be that would be helpful. The transaction file is up to around 9000 after 1 hour 40 mins and as stated before the backup is 8.7 GB. Is it possible to estimate the number of transactions in the backup file? Thanks again for your help. Nic From: Barry Bishop [barry.bis...@ontotext.com] Sent: 13 April 2012 12:33 To: GOULD, Nicholas Cc: Barry Norton; owlim-discussion@ontotext.com Subject: Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. Hello Nicholas, This can take a long time, I'm afraid. If the indices are damaged then OWLIM will replay it's entire transaction log in order to recreate them. My suggestion is just to wait until it completes. Terminating the recovery process means that the next restart will fail to initialise and cause the the message in the subject of this email to be displayed. The good news is that the transaction mechanisms in OWLIM 5.0 have been completely refactored in order to fix these kinds of inconveniences. We will release this version next week. Please see: http://owlim.ontotext.com/display/OWLIMv50/OWLIM-SE+Indexing+Specifics#OWLIM-SEIndexingSpecifics-TransactionMode I hope this helps, barry Barry Bishop OWLIM Product Manager Ontotext AD Tel: +43 650 2000 237 email: barry.bis...@ontotext.com www.ontotext.com On 13/04/12 13:24, GOULD, Nicholas wrote: All, Does anyone have experience of how long OWLIM should take to go through it's transaction log and update all indicies? We have loaded a large volume of data (~7.5 GB of ttl files) into a repository and have a backup file of 8.7GB. We stopped Tomcat and on restart it is now going through transactions, presumably synching any differences between the disk and memory indicies when shutdown. Many thanks, Nic From: owlim-discussion-boun...@ontotext.com [owlim-discussion-boun...@ontotext.com] on behalf of Barry Bishop [barry.bis...@ontotext.com] Sent: 01 April 2012 09:12 To: Barry Norton Cc: owlim-discussion@ontotext.com Subject: Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. Hello all, This message is displayed (and OWLIM refuses to continue) only after the following sequence of events: 1. OWLIM is stopped and not shutdown properly 2. OWLIM is restarted and the recover process starts 3. OWLIM is stopped again without allowing the recovery process to finish 4. OWLIM is restarted At this point, the recovery code is aware of the backed up indices, some half built replacement indices and the transaction log. Without knowing which set of indices is correct, it informs the user and stops. (This is not an ideal situation and it is one that can not occur in OWLIM 5) The important point is that you should always allow OWLIM time to shutdown properly and should not kill tomcat, even if it seems to take a long time. If you have got in to this situation, the probable means to get started again is to shut down tomcat and delete any *.old files in the owlim storage folder. Then restart and allow time for OWLIM to replay its transaction log. The cause of this is that OWLIM 4 (and before) uses a lazy writing strategy for dirty pages. These are only swapped to disk when memory has been exhausted and more pages are required. In other words, the state of the indices on disk and in memory can diverge quite considerably and in the event of a sudden stop, the correct indices must be rebuilt using the transaction log (a file called 'backup' in the OWLIM storage folder). OWLIM 5 still has this mode of operation available (it is called 'fast' mode), but the default mode flushes updates to disk at the point of each commit. It does this in such a way, that a failure at any time leaves the disk image
Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore.
Hi Nicholas, Potentially yes, but if that is the case I am surprised that it is running so slowly. Did you have reasonable performance during loading and how long did that take? barry On 13/04/12 14:32, GOULD, Nicholas wrote: Thanks Barry, Our ttl files are small, but lots of them, only around 30 triples per file. I think, extrapolating from what you said below, that this means we are looking at a very high number of transactions? Nic From: Barry Bishop [barry.bis...@ontotext.com] Sent: 13 April 2012 13:15 To: GOULD, Nicholas Cc: Barry Norton; owlim-discussion@ontotext.com Subject: Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. Hi Nicholas, Yikes, sorry to hear that you are in such a tight spot. Do you know roughly how big your transactions are? This would provide a way to make a guess about how many are in the backup file. I just tried an experiment: LUBM(1) is made up of 15 files of approximately 10,000 statements each. The size of the backup file is 4MB. If you (by some fluke) have the same average transaction size then I would guess there are this many transactions: (8000MB/4MB)*15 = 30,000 transactions. This is very 'back of a fag packet stuff' and depends completely on your average transaction size. If it was 5000 statements then expect 60,000 transactions, if it was 30,000 statements per transaction then expect about 10,000 transactions. Sorry, but this is the best I can do at this stage. Best of luck, barry On 13/04/12 13:47, GOULD, Nicholas wrote: Thanks for the quick response Barry. We figured that we would have to wait for it to complete. I have a project manager breathing down my neck for an ETA on this so if you are able to give us an idea of how long 'a long time' could be that would be helpful. The transaction file is up to around 9000 after 1 hour 40 mins and as stated before the backup is 8.7 GB. Is it possible to estimate the number of transactions in the backup file? Thanks again for your help. Nic From: Barry Bishop [barry.bis...@ontotext.com] Sent: 13 April 2012 12:33 To: GOULD, Nicholas Cc: Barry Norton; owlim-discussion@ontotext.com Subject: Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. Hello Nicholas, This can take a long time, I'm afraid. If the indices are damaged then OWLIM will replay it's entire transaction log in order to recreate them. My suggestion is just to wait until it completes. Terminating the recovery process means that the next restart will fail to initialise and cause the the message in the subject of this email to be displayed. The good news is that the transaction mechanisms in OWLIM 5.0 have been completely refactored in order to fix these kinds of inconveniences. We will release this version next week. Please see: http://owlim.ontotext.com/display/OWLIMv50/OWLIM-SE+Indexing+Specifics#OWLIM-SEIndexingSpecifics-TransactionMode I hope this helps, barry Barry Bishop OWLIM Product Manager Ontotext AD Tel: +43 650 2000 237 email: barry.bis...@ontotext.com www.ontotext.com On 13/04/12 13:24, GOULD, Nicholas wrote: All, Does anyone have experience of how long OWLIM should take to go through it's transaction log and update all indicies? We have loaded a large volume of data (~7.5 GB of ttl files) into a repository and have a backup file of 8.7GB. We stopped Tomcat and on restart it is now going through transactions, presumably synching any differences between the disk and memory indicies when shutdown. Many thanks, Nic From: owlim-discussion-boun...@ontotext.com [owlim-discussion-boun...@ontotext.com] on behalf of Barry Bishop [barry.bis...@ontotext.com] Sent: 01 April 2012 09:12 To: Barry Norton Cc: owlim-discussion@ontotext.com Subject: Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. Hello all, This message is displayed (and OWLIM refuses to continue) only after the following sequence of events: 1. OWLIM is stopped and not shutdown properly 2. OWLIM is restarted and the recover process starts 3. OWLIM is stopped again without allowing the recovery process to finish 4. OWLIM is restarted At this point, the recovery code is aware of the backed up indices, some half built replacement indices and the transaction log. Without knowing which set of indices is correct, it informs the user and stops. (This is not an ideal situation and it is one that can not occur in OWLIM 5) The important point is that you should always allow OWLIM time to shutdown properly and should not kill tomcat, even if it seems to take a long time. If you have got in to this situation, the probable means to get started again is to shut down tomcat and delete any *.old files
Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore.
Load took around 4 hours. From: Barry Bishop [barry.bis...@ontotext.com] Sent: 13 April 2012 13:52 To: GOULD, Nicholas Cc: Barry Norton; owlim-discussion@ontotext.com Subject: Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. Hi Nicholas, Potentially yes, but if that is the case I am surprised that it is running so slowly. Did you have reasonable performance during loading and how long did that take? barry On 13/04/12 14:32, GOULD, Nicholas wrote: Thanks Barry, Our ttl files are small, but lots of them, only around 30 triples per file. I think, extrapolating from what you said below, that this means we are looking at a very high number of transactions? Nic From: Barry Bishop [barry.bis...@ontotext.com] Sent: 13 April 2012 13:15 To: GOULD, Nicholas Cc: Barry Norton; owlim-discussion@ontotext.com Subject: Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. Hi Nicholas, Yikes, sorry to hear that you are in such a tight spot. Do you know roughly how big your transactions are? This would provide a way to make a guess about how many are in the backup file. I just tried an experiment: LUBM(1) is made up of 15 files of approximately 10,000 statements each. The size of the backup file is 4MB. If you (by some fluke) have the same average transaction size then I would guess there are this many transactions: (8000MB/4MB)*15 = 30,000 transactions. This is very 'back of a fag packet stuff' and depends completely on your average transaction size. If it was 5000 statements then expect 60,000 transactions, if it was 30,000 statements per transaction then expect about 10,000 transactions. Sorry, but this is the best I can do at this stage. Best of luck, barry On 13/04/12 13:47, GOULD, Nicholas wrote: Thanks for the quick response Barry. We figured that we would have to wait for it to complete. I have a project manager breathing down my neck for an ETA on this so if you are able to give us an idea of how long 'a long time' could be that would be helpful. The transaction file is up to around 9000 after 1 hour 40 mins and as stated before the backup is 8.7 GB. Is it possible to estimate the number of transactions in the backup file? Thanks again for your help. Nic From: Barry Bishop [barry.bis...@ontotext.com] Sent: 13 April 2012 12:33 To: GOULD, Nicholas Cc: Barry Norton; owlim-discussion@ontotext.com Subject: Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. Hello Nicholas, This can take a long time, I'm afraid. If the indices are damaged then OWLIM will replay it's entire transaction log in order to recreate them. My suggestion is just to wait until it completes. Terminating the recovery process means that the next restart will fail to initialise and cause the the message in the subject of this email to be displayed. The good news is that the transaction mechanisms in OWLIM 5.0 have been completely refactored in order to fix these kinds of inconveniences. We will release this version next week. Please see: http://owlim.ontotext.com/display/OWLIMv50/OWLIM-SE+Indexing+Specifics#OWLIM-SEIndexingSpecifics-TransactionMode I hope this helps, barry Barry Bishop OWLIM Product Manager Ontotext AD Tel: +43 650 2000 237 email: barry.bis...@ontotext.com www.ontotext.com On 13/04/12 13:24, GOULD, Nicholas wrote: All, Does anyone have experience of how long OWLIM should take to go through it's transaction log and update all indicies? We have loaded a large volume of data (~7.5 GB of ttl files) into a repository and have a backup file of 8.7GB. We stopped Tomcat and on restart it is now going through transactions, presumably synching any differences between the disk and memory indicies when shutdown. Many thanks, Nic From: owlim-discussion-boun...@ontotext.com [owlim-discussion-boun...@ontotext.com] on behalf of Barry Bishop [barry.bis...@ontotext.com] Sent: 01 April 2012 09:12 To: Barry Norton Cc: owlim-discussion@ontotext.com Subject: Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. Hello all, This message is displayed (and OWLIM refuses to continue) only after the following sequence of events: 1. OWLIM is stopped and not shutdown properly 2. OWLIM is restarted and the recover process starts 3. OWLIM is stopped again without allowing the recovery process to finish 4. OWLIM is restarted At this point, the recovery code is aware of the backed up indices, some half built replacement indices and the transaction log. Without
Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore.
Hello all, This message is displayed (and OWLIM refuses to continue) only after the following sequence of events: 1. OWLIM is stopped and not shutdown properly 2. OWLIM is restarted and the recover process starts 3. OWLIM is stopped again without allowing the recovery process to finish 4. OWLIM is restarted At this point, the recovery code is aware of the backed up indices, some half built replacement indices and the transaction log. Without knowing which set of indices is correct, it informs the user and stops. (This is not an ideal situation and it is one that can not occur in OWLIM 5) The important point is that you should always allow OWLIM time to shutdown properly and should not kill tomcat, even if it seems to take a long time. If you have got in to this situation, the probable means to get started again is to shut down tomcat and delete any *.old files in the owlim storage folder. Then restart and allow time for OWLIM to replay its transaction log. The cause of this is that OWLIM 4 (and before) uses a lazy writing strategy for dirty pages. These are only swapped to disk when memory has been exhausted and more pages are required. In other words, the state of the indices on disk and in memory can diverge quite considerably and in the event of a sudden stop, the correct indices must be rebuilt using the transaction log (a file called 'backup' in the OWLIM storage folder). OWLIM 5 still has this mode of operation available (it is called 'fast' mode), but the default mode flushes updates to disk at the point of each commit. It does this in such a way, that a failure at any time leaves the disk image consistent with the state of the database at the end of the last successful commit operation. For sustained loading operations, this can have lower performance, but the advantages are: 1. instant shutdown time 2. instant recovery time (even if the last instance was stopped abruptly) 3. improved concurrency of reads and write operations This last point means that export the database programmatically is now a viable means to backup the database. To summarise: OWLIM 4 - this error occurs during initialisation when the OWLIM process has been killed, restarted and not allowed time to recover properly, i.e. killed again. It can be avoided by giving owlim enough time to shutdown properly (which can be a long time after loading large amounts of data on a machine with large memory). If it does happen, shutdown, delete the *.old files, restart and wait. OWLIM 5 - using the default transaction mode, this problem will not occur. The fast transaction mode is available when administrators need maximum performance and reduced query/update concurrency, but will have no guarantees about data security if OWLIM is killed before shutting properly - switching between transaction modes requires a restart. I hope this helps, barry Barry Bishop OWLIM Product Manager Ontotext AD Tel: +43 650 2000 237 email: barry.bis...@ontotext.com www.ontotext.com On 01/04/12 01:29, Barry Norton wrote: Guys, how are you shutting down your database that it needs to be recovered so often? And are you interrupting the recovery process? Barry On 31/03/2012 20:44, Blasetti, Luciano (CIOK) wrote: Hi All, I also experienced quite frequently this issue on a tomcat instance running only openrdf-sesame (OWLIM 4.3 SE) and openrdf-workbench. No way to restore the database. Moreover, since there's no out-of-the-box functionalities to perform a hot backup on a non clustered configuration (and the programmatic export can't work on large triplestore due to timeout issues), this strange behavior leads to data loss. I hope that the forthcoming OWLIM 5.0 will be more robust and reliable. Regards, Luciano From: owlim-discussion-boun...@ontotext.com [owlim-discussion-boun...@ontotext.com] On Behalf Of Marek Šurek [marek_su...@yahoo.co.uk] Sent: Saturday, March 31, 2012 11:18 AM To: Monika Solanki; owlim-discussion@ontotext.com Subject: Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. Hi Monika, I also experienced this exception number of times on OWLIM 4.3. The problem appears only when I ran openrdf-sesame and client application at the same Tomcat server and due to developing client application I had to restart Tomcat very offen to see changes in my client application. My solution was to divide those two applications into different tomcat instances so now they run independently(my client application runs on different Tomcat as openrdf-sesame) and I don't restart Tomcat instance with openrdf-sesame.war almost never. I hope this would help, Marek From: Monika Solankimsolanki.maili...@gmail.com To: owlim-discussion@ontotext.com Sent: Saturday, 31 March 2012, 10:59 Subject: [Owlim-discussion] CRITICAL FAILURE !!! The database
[Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore.
Hello Again, A couple of times I have had this error thrown to me and I have had to populate the repository again. CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. I would appreciate any help on how I can possibly avoid this error. Thanks, Monika ___ Owlim-discussion mailing list Owlim-discussion@ontotext.com http://ontomail.semdata.org/cgi-bin/mailman/listinfo/owlim-discussion
Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore.
Hi Monika, I also experienced this exception number of times on OWLIM 4.3. The problem appears only when I ran openrdf-sesame and client application at the same Tomcat server and due to developing client application I had to restart Tomcat very offen to see changes in my client application. My solution was to divide those two applications into different tomcat instances so now they run independently(my client application runs on different Tomcat as openrdf-sesame) and I don't restart Tomcat instance with openrdf-sesame.war almost never. I hope this would help, Marek From: Monika Solanki msolanki.maili...@gmail.com To: owlim-discussion@ontotext.com Sent: Saturday, 31 March 2012, 10:59 Subject: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. Hello Again, A couple of times I have had this error thrown to me and I have had to populate the repository again. CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. I would appreciate any help on how I can possibly avoid this error. Thanks, Monika ___ Owlim-discussion mailing list Owlim-discussion@ontotext.com http://ontomail.semdata.org/cgi-bin/mailman/listinfo/owlim-discussion___ Owlim-discussion mailing list Owlim-discussion@ontotext.com http://ontomail.semdata.org/cgi-bin/mailman/listinfo/owlim-discussion
Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore.
Hi All, I also experienced quite frequently this issue on a tomcat instance running only openrdf-sesame (OWLIM 4.3 SE) and openrdf-workbench. No way to restore the database. Moreover, since there's no out-of-the-box functionalities to perform a hot backup on a non clustered configuration (and the programmatic export can't work on large triplestore due to timeout issues), this strange behavior leads to data loss. I hope that the forthcoming OWLIM 5.0 will be more robust and reliable. Regards, Luciano From: owlim-discussion-boun...@ontotext.com [owlim-discussion-boun...@ontotext.com] On Behalf Of Marek Šurek [marek_su...@yahoo.co.uk] Sent: Saturday, March 31, 2012 11:18 AM To: Monika Solanki; owlim-discussion@ontotext.com Subject: Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. Hi Monika, I also experienced this exception number of times on OWLIM 4.3. The problem appears only when I ran openrdf-sesame and client application at the same Tomcat server and due to developing client application I had to restart Tomcat very offen to see changes in my client application. My solution was to divide those two applications into different tomcat instances so now they run independently(my client application runs on different Tomcat as openrdf-sesame) and I don't restart Tomcat instance with openrdf-sesame.war almost never. I hope this would help, Marek From: Monika Solanki msolanki.maili...@gmail.com To: owlim-discussion@ontotext.com Sent: Saturday, 31 March 2012, 10:59 Subject: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. Hello Again, A couple of times I have had this error thrown to me and I have had to populate the repository again. CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. I would appreciate any help on how I can possibly avoid this error. Thanks, Monika ___ Owlim-discussion mailing list Owlim-discussion@ontotext.commailto:Owlim-discussion@ontotext.com http://ontomail.semdata.org/cgi-bin/mailman/listinfo/owlim-discussion ___ Owlim-discussion mailing list Owlim-discussion@ontotext.com http://ontomail.semdata.org/cgi-bin/mailman/listinfo/owlim-discussion
Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore.
Guys, how are you shutting down your database that it needs to be recovered so often? And are you interrupting the recovery process? Barry On 31/03/2012 20:44, Blasetti, Luciano (CIOK) wrote: Hi All, I also experienced quite frequently this issue on a tomcat instance running only openrdf-sesame (OWLIM 4.3 SE) and openrdf-workbench. No way to restore the database. Moreover, since there's no out-of-the-box functionalities to perform a hot backup on a non clustered configuration (and the programmatic export can't work on large triplestore due to timeout issues), this strange behavior leads to data loss. I hope that the forthcoming OWLIM 5.0 will be more robust and reliable. Regards, Luciano From: owlim-discussion-boun...@ontotext.com [owlim-discussion-boun...@ontotext.com] On Behalf Of Marek Šurek [marek_su...@yahoo.co.uk] Sent: Saturday, March 31, 2012 11:18 AM To: Monika Solanki; owlim-discussion@ontotext.com Subject: Re: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. Hi Monika, I also experienced this exception number of times on OWLIM 4.3. The problem appears only when I ran openrdf-sesame and client application at the same Tomcat server and due to developing client application I had to restart Tomcat very offen to see changes in my client application. My solution was to divide those two applications into different tomcat instances so now they run independently(my client application runs on different Tomcat as openrdf-sesame) and I don't restart Tomcat instance with openrdf-sesame.war almost never. I hope this would help, Marek From: Monika Solankimsolanki.maili...@gmail.com To: owlim-discussion@ontotext.com Sent: Saturday, 31 March 2012, 10:59 Subject: [Owlim-discussion] CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. Hello Again, A couple of times I have had this error thrown to me and I have had to populate the repository again. CRITICAL FAILURE !!! The database restorer tool has already failed once before to perform image restore. I would appreciate any help on how I can possibly avoid this error. Thanks, Monika ___ Owlim-discussion mailing list Owlim-discussion@ontotext.commailto:Owlim-discussion@ontotext.com http://ontomail.semdata.org/cgi-bin/mailman/listinfo/owlim-discussion ___ Owlim-discussion mailing list Owlim-discussion@ontotext.com http://ontomail.semdata.org/cgi-bin/mailman/listinfo/owlim-discussion ___ Owlim-discussion mailing list Owlim-discussion@ontotext.com http://ontomail.semdata.org/cgi-bin/mailman/listinfo/owlim-discussion