Jump to content
offthewall

Retrospect 10 performance intolerable

Recommended Posts

I just upgraded from Retrospect 8 to 10. While 8 was no picnic, 10 is absurd. Performance in the console is SUPER slow, buggy and unresponsive both working locally on the server or over the network. I'm running on an OS Server 10.6.8 Xserve. 2x2.26 Quad Core Xenon with a light load of file sharing to 5-10 clients.

 

Does anyone actually get backups done with this expensive garbage?

  • Like 1

Share this post


Link to post
Share on other sites

Quite frustrating. I've been working with this software all morning. I've now completely uninstalled Retrospect on the server and re-installed. I can't go 5 minutes in the console either locally or over the network (macbook air 10.8.3) with out it freezing completely up. Moving from one function to the next results in endless spinning beach balls.

 

At this stage I'm simply trying to re-establish media sets and log clients in. I can't get a configuration working to even start to back up.

 

I've been working with the various versions of Retro 8 since 8.0, so I know the basics, this version is just SO slow and buggy, I can't get it running.

 

HELP!!!!!!!!!!!

Share this post


Link to post
Share on other sites

Sounds odd. I have seen improvements 8 ->9 -> 10 -> 10.1, for the most part. There may be something bad going on.

 

The protocol between the console and the engine is "chatty". What is the connection you are using, and how's the latency? (ping time)

Share this post


Link to post
Share on other sites
I can't go 5 minutes in the console either locally or over the network (macbook air 10.8.3) with out it freezing completely up.

 

 

Is this what you saw upon initial install? Before you created/located Media Sets? Before you logged in Sources?

 

 

Where are your Catalog files stored?

What are your sources for Backup?

What type of Media Sets are you using?

What devices are connected to the XServe?

Share this post


Link to post
Share on other sites

Sounds odd. I have seen improvements 8 ->9 -> 10 -> 10.1, for the most part. There may be something bad going on.

 

The protocol between the console and the engine is "chatty". What is the connection you are using, and how's the latency? (ping time)

 

The performance issues is seen both working on the server machine locally as well as over the local gigabit wired network. Here is the ping time to the server over ethernet.

 

 

Ping has started…

PING pikserver.com (192.168.1.213): 56 data bytes

64 bytes from 192.168.1.213: icmp_seq=0 ttl=64 time=0.035 ms

64 bytes from 192.168.1.213: icmp_seq=1 ttl=64 time=0.087 ms

64 bytes from 192.168.1.213: icmp_seq=2 ttl=64 time=0.104 ms

64 bytes from 192.168.1.213: icmp_seq=3 ttl=64 time=0.091 ms

64 bytes from 192.168.1.213: icmp_seq=4 ttl=64 time=0.086 ms

64 bytes from 192.168.1.213: icmp_seq=5 ttl=64 time=0.094 ms

64 bytes from 192.168.1.213: icmp_seq=6 ttl=64 time=0.090 ms

64 bytes from 192.168.1.213: icmp_seq=7 ttl=64 time=0.090 ms

64 bytes from 192.168.1.213: icmp_seq=8 ttl=64 time=0.092 ms

64 bytes from 192.168.1.213: icmp_seq=9 ttl=64 time=0.087 ms

--- 192.168.1.213 ping statistics ---

10 packets transmitted, 10 packets received, 0.0% packet loss

round-trip min/avg/max/stddev = 0.035/0.086/0.104/0.018 ms

Share this post


Link to post
Share on other sites

If you see the performance problem with a local console, that eliminates the network performance. (as long as you are connecting to 127.0.0.1 - you might want to check) That means that whatever is dragging is local to the machine, unless the catalogs or config files are "remote" for some reason, like mounted NFS or AFS.

 

I wish I were in NYC. It might be interesting to do some analysis to see what's so slow. All I can tell you is that it's not a problem for me. There must be something peculiar in your installation, and something that changed from 8.0 to 10.1.

 

One question: Do you have any heavy-duty debugging settings turned on in Retrospect? I think those could cause performance problems if set aggressively enough.

Share this post


Link to post
Share on other sites

I likewise have had terrible lag times after moving from 10.0 to 10.1. I'm running on OSX 10.8.2 on an xserver 2.26 Quad-core Intel with 6 Gb of RAM. Only running AFP services on that machine and after upgrading to 10.1 I can barely use retrospect.

Share this post


Link to post
Share on other sites

If you see the performance problem with a local console, that eliminates the network performance. (as long as you are connecting to 127.0.0.1 - you might want to check) That means that whatever is dragging is local to the machine, unless the catalogs or config files are "remote" for some reason, like mounted NFS or AFS.

 

I wish I were in NYC. It might be interesting to do some analysis to see what's so slow. All I can tell you is that it's not a problem for me. There must be something peculiar in your installation, and something that changed from 8.0 to 10.1.

 

One question: Do you have any heavy-duty debugging settings turned on in Retrospect? I think those could cause performance problems if set aggressively enough.

 

I'm not sure, not intentionally. Where would I check for those debugging settings?

 

I get spurts of responsiveness. For about 10-15 seconds, I can zip about the UI, then spinning ball again.

Share this post


Link to post
Share on other sites

I'm not sure, not intentionally. Where would I check for those debugging settings?

...

I don't think there's any way of checking/changing them except by going to the config file directly. It's in /Library/Application Support/Retrospect/RetrospectEngine.bundle/Contents/MacOS/retro.ini, but if you don't know where it is, you probably haven't changed it :) In the finder you have to use the Finder context menu Show Package Contents to see inside RetrospectEngine.bundle. retro.ini is a plain text file.

Share this post


Link to post
Share on other sites

I can't go 5 minutes in the console either locally or over the network (macbook air 10.8.3) with out it freezing completely up. Moving from one function to the next results in endless spinning beach balls.

Do you really mean freezing, or do you mean frequent SPODs/beachballs in the console app?

 

We've been experiencing a lot of SPODs with the Retro 10.1(221) console. Most are of just a few seconds' duration, but if the Retrospect engine is busy for any reason, they can last for 10 seconds or more. I've been attributing this to the way the console now updates the data from the engine more frequently. I've found them to be annoying but not paralyzing, though the delays can be rather frustrating when all you want to do is make a quick change or view some information. Given how the earlier versions of the console used to be rather haphazard in the way they updated themselves from the engine, I had been thinking that the tradeoff between responsiveness and timely data was worth it, but now I'm wondering why we can't have both.

 

I'm curious what kinds of configurations those of us have who are experiencing reduced console responsiveness. It would seem logical that a more complex configuration would require more frequent polling of the engine by the console to see if changes have been made.

 

Our setup consists of 42 local and client sources, most of which have between two and seven favorite folders; 6 media sets, two of which have 14 members each; 21 scripts; and 12 rules, a couple of which have around 30 elements. 1,800 past backups are listed. This doesn't seem particularly complex to me. Your thoughts?

Share this post


Link to post
Share on other sites

I find that multiple concurrent proactive activities can cause SPODs -- particularly when getting to the "closing" state of a client backup. That's fairly consistent for me when generating SPODs

Share this post


Link to post
Share on other sites

Do you really mean freezing, or do you mean frequent SPODs/beachballs in the console app?

 

We've been experiencing a lot of SPODs with the Retro 10.1(221) console. Most are of just a few seconds' duration, but if the Retrospect engine is busy for any reason, they can last for 10 seconds or more. I've been attributing this to the way the console now updates the data from the engine more frequently. I've found them to be annoying but not paralyzing, though the delays can be rather frustrating when all you want to do is make a quick change or view some information. Given how the earlier versions of the console used to be rather haphazard in the way they updated themselves from the engine, I had been thinking that the tradeoff between responsiveness and timely data was worth it, but now I'm wondering why we can't have both.

 

I'm curious what kinds of configurations those of us have who are experiencing reduced console responsiveness. It would seem logical that a more complex configuration would require more frequent polling of the engine by the console to see if changes have been made.

 

Our setup consists of 42 local and client sources, most of which have between two and seven favorite folders; 6 media sets, two of which have 14 members each; 21 scripts; and 12 rules, a couple of which have around 30 elements. 1,800 past backups are listed. This doesn't seem particularly complex to me. Your thoughts?

 

Yes, your description of SPODs is what I mean by freezing. The problem is that the 10 second or more SPOD is followed by only a 1 or 2 sec period of responsiveness then the beach ball returns. It took me almost an hour to change the media set on 10 scripts this afternoon. You get that second or two to try to click what's needed and then you wind up clicking something that your were trying to click 15 minutes ago.

 

Underneath, the backups are continuing at a decently fast pace.

Share this post


Link to post
Share on other sites

I now have two completely different Retrospect installations experiencing the same horrible performance in the Console. Separate companies, different locations. The only thing in common is they are both running on an XServe with 10.6.8 OSX Server OS. This installation was also running V8 without this terrible lag.

 

Can anyone at Retrospect comment on this?

Share this post


Link to post
Share on other sites

Very sorry for not replying sooner.

 

Dennis, thank you for uploading the Activity Monitor sample of the SPOD'ing Mac console. We think it pinpoints your issue. It's a known problem (#3798) with 10.1.0.221 that comes up in certain environments, and it will be fixed in the next update.

 

For everyone else seeing seeings SPODs or sporadic hanging, it would be a great help if you could also sample the hung Retrospect process with Activity Monitor and upload the resulting file to a support ticket. That way, we can make sure we fix all instances of these hangs in the next update. As iCompute suggested, the root cause is the chatty console/engine protocol that ensures the console information is up-to-date.

 

Please contact our support team if you are interested in seeing if an upcoming test build will fix your issue.

Share this post


Link to post
Share on other sites

Very sorry for not replying sooner.

 

Dennis, thank you for uploading the Activity Monitor sample of the SPOD'ing Mac console. We think it pinpoints your issue. It's a known problem (#3798) with 10.1.0.221 that comes up in certain environments, and it will be fixed in the next update.

 

For everyone else seeing seeings SPODs or sporadic hanging, it would be a great help if you could also sample the hung Retrospect process with Activity Monitor and upload the resulting file to a support ticket. That way, we can make sure we fix all instances of these hangs in the next update. As iCompute suggested, the root cause is the chatty console/engine protocol that ensures the console information is up-to-date.

 

Please contact our support team if you are interested in seeing if an upcoming test build will fix your issue.

 

I would definitely like to see a test build addressing this issue.

Share this post


Link to post
Share on other sites

The more RAM you have seems to ameliorate the prob big time. Ran Mac Pro w/8GB & 8/9/10/10.1 and just recently went to 16GB - big difference in peerformance.

 

If you ( during SPOD ) <force quit> and look at Retro you'll see it shows [ Retrospect Not Responding ], and if you wait a bit it resumes itself pretty soon. Still...

 

Could manage resources ( if that's the problem area ) better ....

Share this post


Link to post
Share on other sites

I just updated Retrospect from v9 to 10. While v9 ran smoothly in the background, v10 hogs my CPU like crazy and throws these errors into console:

 

06.05.13 17:52:45,573 Retrospect[44513]: BackupSet::setBackupSet exception: <RefBackupset::GetSnapshotsFiltered> {MHD2_andy} Other access failure(errcode = -1101)

06.05.13 17:52:45,576 Retrospect[44513]: BackupSet::setBackupSet exception: <RefBackupset::GetMembers> {MHD2_andy} Other access failure(errcode = -1101)

 

 

iMac 27" mid 2010, 10.8.3, 24GB RAM

 

Share this post


Link to post
Share on other sites

Which Retrospect v10, specifically? Versions earlier than 10.1.0 have problems with Instant Scan burning large amounts of CPU time.

Share this post


Link to post
Share on other sites

Which Retrospect v10, specifically? Versions earlier than 10.1.0 have problems with Instant Scan burning large amounts of CPU time.

 

It's the latest Version 10.1.0 (221). All updates done. The issue is at the computer running the console not the engine. No client installed on the console machine. The freezes are in the console app.

Share this post


Link to post
Share on other sites

prl asked:

 

Which Retrospect v10, specifically?

 

To which Dennis responded:

 

 

It's the latest v10. All updates done.

 

This thread will live forever, and prl would have to cross-check the date of your post against the release dates of the software builds just to get the answer to his question.

 

Be a lot easier if you just said which version of Retrospect v10, specifically...

 

-dave

Share this post


Link to post
Share on other sites

Also, I was responding to tiptronic, not to Dennis, so it's actually tiptronic's Retrospect version number I was asking about, not Dennis's.

Share this post


Link to post
Share on other sites

I would definitely like to see a test build addressing this issue.

 

 

Very sorry for not replying sooner.

 

Dennis, thank you for uploading the Activity Monitor sample of the SPOD'ing Mac console. We think it pinpoints your issue. It's a known problem (#3798) with 10.1.0.221 that comes up in certain environments, and it will be fixed in the next update.

 

For everyone else seeing seeings SPODs or sporadic hanging, it would be a great help if you could also sample the hung Retrospect process with Activity Monitor and upload the resulting file to a support ticket. That way, we can make sure we fix all instances of these hangs in the next update. As iCompute suggested, the root cause is the chatty console/engine protocol that ensures the console information is up-to-date.

 

Please contact our support team if you are interested in seeing if an upcoming test build will fix your issue.

Is there a test build or update available yet? I would LOVE to test it for you. The constant SPOD'ing is ever present with both of my installations. thx.

Share this post


Link to post
Share on other sites

Is there a test build or update available yet? I would LOVE to test it for you. The constant SPOD'ing is ever present with both of my installations. thx.

I have been running the pre-release V 10.2.0.196 for a couple of weeks now and Retrospect 10 performance is now back on par with V9 (a good thing). It is usable again for me on my two OSX 10.6.8 Servers.

Share this post


Link to post
Share on other sites

I just upgraded from Retrospect 8 to 10. While 8 was no picnic, 10 is absurd. Performance in the console is SUPER slow, buggy and unresponsive both working locally on the server or over the network. I'm running on an OS Server 10.6.8 Xserve. 2x2.26 Quad Core Xenon with a light load of file sharing to 5-10 clients.

I have a similar machine (although a Mac Pro instead of an XServe) backing up 50 clients and do not see SPODs on the local or remote console. The console is a bit sluggish (understatement), but is not hanging - once it's completed the initial sync it works reasonably well. How much RAM does it have? (and did 10.2 resolve the issue for you?)

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×