Re: planner timeout
René Kanters wrote: Hi Paul, thanks for getting back to me on this. I don't think it is firewall issues because up till now the backup ran fine. Actually it ran fine last night. Nobody was doing any big calculations on it, which could have affected that. Also, so maybe something else strange on the network was going on. Is there a way to affect the 30 sec timeout for the planner's ACK timeout? It's hardcoded. Anyway, no reasonable program should be expected to send a simple ACK-packet back taking more than 30 seconds :-) If it can't then anything really bad is wrong, like crashing client programs, firewalls, routing problems etc. Cheers, René On Nov 29, 2007, at 10:24 AM, Paul Bijnens wrote: On 2007-11-27 18:13, René Kanters wrote: Hi, I have been running into problems that some of my systems are heavily used for long computations making them somewhat less responsive. Last night I ran into the issue that four systems did not send acknowledgments back to the dumper on time during the planning process: planner: ERROR Request to werner.richmond.edu failed: timeout waiting for ACK I looked into allowing more time for that stage, which I believe etimeout should allow, but my amanda.conf has 'etimeout 600' in it while the planner's debug file ends with: security_seterror(handle=0x3038a0, driver=0xa2a0c (BSD) error=timeout waiting for ACK) security_close(handle=0x3038a0, driver=0xa2a0c (BSD)) planner: time 29.898: pid 3734 finish time Tue Nov 27 00:45:36 2007 suggesting that it still only waits for 30 seconds. planner sends a packet to the client(s) and it expects at least an UDP ACK-packet back within 30 seconds, indicating that the client did receive at least the request. The etimeout is the time that planner will wait for the packet with the different size estimates from the client, which will usually take more than 30 seconds. Am I setting the wrong timeout? So it seems you can't even get a reply back. Firewall issues? -- Paul Bijnens, xplanation Technology ServicesTel +32 16 397.511 Technologielaan 21 bus 2, B-3001 Leuven, BELGIUMFax +32 16 397.512 http://www.xplanation.com/ email: [EMAIL PROTECTED] *** * I think I've got the hang of it now: exit, ^D, ^C, ^\, ^Z, ^Q, ^^, * * F6, quit, ZZ, :q, :q!, M-Z, ^X^C, logoff, logout, close, bye, /bye, * * stop, end, F3, ~., ^]c, +++ ATH, disconnect, halt, abort, hangup, * * PF4, F20, ^X^X, :D::D, KJOB, F14-f-e, F8-e, kill -1 $$, shutdown, * * init 0, kill -9 1, Alt-F4, Ctrl-Alt-Del, AltGr-NumLock, Stop-A, ... * * ... "Are you sure?" ... YES ... Phew ... I'm out * ***
Re: planner timeout
Hi Paul, thanks for getting back to me on this. I don't think it is firewall issues because up till now the backup ran fine. Actually it ran fine last night. Nobody was doing any big calculations on it, which could have affected that. Also, so maybe something else strange on the network was going on. Is there a way to affect the 30 sec timeout for the planner's ACK timeout? Cheers, René On Nov 29, 2007, at 10:24 AM, Paul Bijnens wrote: On 2007-11-27 18:13, René Kanters wrote: Hi, I have been running into problems that some of my systems are heavily used for long computations making them somewhat less responsive. Last night I ran into the issue that four systems did not send acknowledgments back to the dumper on time during the planning process: planner: ERROR Request to werner.richmond.edu failed: timeout waiting for ACK I looked into allowing more time for that stage, which I believe etimeout should allow, but my amanda.conf has 'etimeout 600' in it while the planner's debug file ends with: security_seterror(handle=0x3038a0, driver=0xa2a0c (BSD) error=timeout waiting for ACK) security_close(handle=0x3038a0, driver=0xa2a0c (BSD)) planner: time 29.898: pid 3734 finish time Tue Nov 27 00:45:36 2007 suggesting that it still only waits for 30 seconds. planner sends a packet to the client(s) and it expects at least an UDP ACK-packet back within 30 seconds, indicating that the client did receive at least the request. The etimeout is the time that planner will wait for the packet with the different size estimates from the client, which will usually take more than 30 seconds. Am I setting the wrong timeout? So it seems you can't even get a reply back. Firewall issues? -- Paul Bijnens, xplanation Technology ServicesTel +32 16 397.511 Technologielaan 21 bus 2, B-3001 Leuven, BELGIUMFax +32 16 397.512 http://www.xplanation.com/ email: [EMAIL PROTECTED] ** * * I think I've got the hang of it now: exit, ^D, ^C, ^\, ^Z, ^Q, ^^, * * F6, quit, ZZ, :q, :q!, M-Z, ^X^C, logoff, logout, close, bye, / bye, * * stop, end, F3, ~., ^]c, +++ ATH, disconnect, halt, abort, hangup, * * PF4, F20, ^X^X, :D::D, KJOB, F14-f-e, F8-e, kill -1 $$, shutdown, * * init 0, kill -9 1, Alt-F4, Ctrl-Alt-Del, AltGr-NumLock, Stop- A, ... * * ... "Are you sure?" ... YES ... Phew ... I'm out * ** *
Re: planner timeout
On 2007-11-27 18:13, René Kanters wrote: Hi, I have been running into problems that some of my systems are heavily used for long computations making them somewhat less responsive. Last night I ran into the issue that four systems did not send acknowledgments back to the dumper on time during the planning process: planner: ERROR Request to werner.richmond.edu failed: timeout waiting for ACK I looked into allowing more time for that stage, which I believe etimeout should allow, but my amanda.conf has 'etimeout 600' in it while the planner's debug file ends with: security_seterror(handle=0x3038a0, driver=0xa2a0c (BSD) error=timeout waiting for ACK) security_close(handle=0x3038a0, driver=0xa2a0c (BSD)) planner: time 29.898: pid 3734 finish time Tue Nov 27 00:45:36 2007 suggesting that it still only waits for 30 seconds. planner sends a packet to the client(s) and it expects at least an UDP ACK-packet back within 30 seconds, indicating that the client did receive at least the request. The etimeout is the time that planner will wait for the packet with the different size estimates from the client, which will usually take more than 30 seconds. Am I setting the wrong timeout? So it seems you can't even get a reply back. Firewall issues? -- Paul Bijnens, xplanation Technology ServicesTel +32 16 397.511 Technologielaan 21 bus 2, B-3001 Leuven, BELGIUMFax +32 16 397.512 http://www.xplanation.com/ email: [EMAIL PROTECTED] *** * I think I've got the hang of it now: exit, ^D, ^C, ^\, ^Z, ^Q, ^^, * * F6, quit, ZZ, :q, :q!, M-Z, ^X^C, logoff, logout, close, bye, /bye, * * stop, end, F3, ~., ^]c, +++ ATH, disconnect, halt, abort, hangup, * * PF4, F20, ^X^X, :D::D, KJOB, F14-f-e, F8-e, kill -1 $$, shutdown, * * init 0, kill -9 1, Alt-F4, Ctrl-Alt-Del, AltGr-NumLock, Stop-A, ... * * ... "Are you sure?" ... YES ... Phew ... I'm out * ***
Re: cleaning tapes and integration into Amanda backup scheme?
Craig Dewick wrote: On Mon, 22 Oct 2007, Olivier Nicole wrote: My Sun L9 array has told me it needs a cleaning tape run. I have one so that's no problem but what I'd like to know is if there is a way that Amanda can receive info from the tape drive about the requirement for cleaning and co-ordinate cleaning tape runs as part of the overall backup stragegy? My strategy, all human based, is to have the cleaning tape on the pile of the next set of 6 tapes to be used. I have my tapes pool divided into 3 sets of 6, once I have run through the current set, I move the stack to the back and come up with a new stack, on top of which is the cleaning tape. My L9 tape array has DLT-4 tapes in the first 8 slots, and today I've put a brand new cleaning tape into the 9th slot. There's nothing in my tape server's amanda.conf file relating to cleaning tapes, so I'm wondering where in the Amanda config schema the info about location of cleaning tapes needs to be. Does Amanda itself need to know, or does mtx need to know directly? I have a 16 slot library. I configured Amanda daily backups to work with slots 1-15, so it would never look at slot 16. Originally, I had the cleaning tape in 16, but I've only had to use it once in a year, and that was when I had a faulty tape get caught in the drive. So, I've just removed it, and I use the 16th slot for archive runs and other special cases. From just scanning the wiki with google, it appears that the chg scripts have the capacity to call a cleaning tape, and that can be defined for the scripts, but it isn't built into Amanda per se. With modern drives, it shouldn't be needed much and my inclination is that I would like to be in manual control of it. If the drive is misbehaving and seems to need cleaning, I don't want an automated process to keep it out of sight (out of mind) until it fails and needs major work. That's just my opinion. --- Chris Hoogendyk - O__ Systems Administrator c/ /'_ --- Biology & Geology Departments (*) \(*) -- 140 Morrill Science Center ~~ - University of Massachusetts, Amherst <[EMAIL PROTECTED]> --- Erdös 4